A half-empty glass and the joy of “failing forward”

Guest blog by Silverio Zebral Filho

This is a blog series written by the alumni of the Implementing Public Policy Executive Education Program at the Harvard Kennedy School. Participants successfully completed this 6-month online learning course in December 2020. These are their learning journey stories.

I’ve started my PDIA journey 6 months ago, interested in gaining a deeper learning about alternative approaches to tackle a wicked (ill-defined, multi-sourced, technically complex and politically sensitive) problem in the context of large institutional divergence (weak rules, strong social norms), lack of state capacity, declining interpersonal trust and confidence in government, gradually leading to social decohesion and violence – all boosted by the presence of transnational crime in several societal domains. 

Our challenge was to help The Office of the Presidency of Government of Honduras in improving transparency and accountability in country public sector, targeting a reduction of 20% on corruption victimization measured by the “control of corruption” indicator of MCC BSC Honduras FY2021. This reduction was instrumental to qualify Honduras to apply for a second Millennium Challenge Corporation (MCC) award starting in 2021, estimated in USD 255 millions.

Our role was to facilitate the establishment and put-in-motion a public-private partnership politically autonomous deliberative/executive unit called “Unidad de Incidencia en Políticas Públicas (UIPP)”. UIPP would be a forum to convene participation of non-State actors invited to contribute to agenda-setting and collective action toward the target. Several actors will hold a seat in the UIPP, with specific agents representing a wide set of stakeholders in a sort of “halocracy” format.  

Successful PDIA experiences in other similar developing countries in Africa and Latin-America impels our team to better understand PDIA basics and pilot some PDIA components into a more traditional PC approach that characterizes technical assistance programs funded by international donors to establish delivering units. The P-D-I-A acronym itself was attractive. Delivery units are (or at least might be…) driven (“D”) by specific policy problems (P) to be tackled by solutions emerging from multi-stakeholder interaction (“I”) carried out in some form experimental, gradual, adaptative approach (“A”).      

PDIA method was introduced in June 2020 during the early phases of the initiative in order to better (re)define corruption mitigation along victimization and perception lines and broad initiative scope. Several entry-points were defined in collaboration with several stakeholder and ideation, test, evaluation, adaptation, trial and learn cycles were applied in 2 major entry-points: public procurement (prone to rent-seeking) and civil service (prone to clientelism and patronage).   

Considering the political sensitiveness of the subject, functional progress is well below the line up to date. Progress delay cannot be attributed to PDIA method. Gradual deterioration of political environment within the country and the surge of a scandal that trigged a major replacement of the Big P authorizer and local convener in charge of keeping the initiative alive. This “limbo” situation was just solved in mid-November and the initiative will restart in January 2021. Nevertheless, slow functional progress was by far compensated by progress in legitimacy and supporting and – maybe more important – by learning regarding re-scoping intervention to narrowing entry-points and refocus team energy in promising early-victory domains.  

Regarding progress learning, insights were many and are still coming. PDIA is tremendously useful when it comes to complex policy challenges but must be combined with some degree of a flexible (traditional) PC approach regarding progress along functional axis to comply with log frame and metrics when the implementation is funded by international donors. PDIA reveals itself particularly hard when there is a specific singular metric to be affected, when fiscal restraint implies in limitation to trial-and-error. On the other hand, is easier to use PDIA when is facilitated by an external agent due to reputational gains and credentialism. Multi-agent leadership under polycentric governance in “halocracy” shape is easier to say than done. Collective action coordination is difficult, attribution of responsibility becomes unclear, accountability becomes diffuse. Principal-agent problems, free-riding and opportunism are big problems when benefits are collectively appropriated and merit recognition is difficult to individualize, disincentivizing risk-taking. Problem deconstruction must be managed with caution: few perspectives can lead to stigmatization and/or problem misidentification; too many nuances can lead you to disperse energy along too many entry points without being sure of consequential with the functional progress toward the metric. At the end of the day, the collinearity among entry-points and causality between selected entry-points and problem solution will be always tricky to understand prior to action. 

Interactions by systematic feedback loop are key to learn in uncertain contexts but learn and leads are not enough as progress’ indicators. However, learn and leads have very limited and decreasing marginal effects over (re)authorization. PDIA must often deliver small victories towards functional metrics that can be made visible in order to build public support and “feelgood” effects.   

Strategic communication is key for any interactive methods such as PDIA: narrative crafting, framing and emotional storytelling are “must-use” to connect, mobilize and convince authorizers and team members in a way that can not only buy-in the initiative, but mainly to sell-it (or save face) to the press and the general public. Listening and inquiring using revealing questions is always better than advocacy. Advocacy is always better when it made by somebody else. Emotions are integral part of listening and inquiring and must be brought into the room.

Leadership is a crucial element to manage not only team tasks, but – mainly – team emotion. That depends on: (a) manage people, perceptions, projections and process under a psychological safety environment, (b) build trust by projecting authenticity, empathy and logic and (c) delegating not only responsibilities but resources to comply with them.   

Despite the harsh conditions, these insights were very useful in improving implementation dynamics in our policy challenge in Honduras. Our team regard PDIA method not only as useful but also as “funny” to put in motion: a secret source to break the rigidity of PC traditional approach. In this sense, team aims to move forward with PDIA method, systematize it, applying it to concrete policy challenges and mentoring/coaching colleagues interested in learning the practice. On the practice improvement side, our team missed some guidance on failure (i.e., when a policy challenge fails and it’s time to stop?) and types of failure (i.e political fatigue? fiscal exhaustion? low morale?). PDIA might be improved with an add-on to help policy implementers to identify dead ends, points of no returns and policy challenge failures, providing a framework to learn from the “final failure” when that turns to be the case. Finally, integrating Public Value Accounts (PVA) and PDIA in a joint method would offer a PVA macro-purpose for PDIA micro-actions.       

Despite of some minor limitations, PDIA has a great potential to be applied not only in public sector, but also in private sector. Some elements such as problem construction/deconstruction and systematic interaction in short feedback loops are already in use in other domains as software (agile) and product (design thinking) development. PDIA techniques are transferable to user-centered open innovation protocols in private sector. Maybe more important, PDIA help to bring soft leadership skills (adaptative leadership, active listening, communication, negotiation, teaming, “muddling through” innovation, scaling by adaptation) back into the center of organizational and management sciences. Finally, PDIA is aligned with the current “stateof-art” of policy studies: peer-collaboration, open polycentric governance, policy complexity, behavioral-informed policy making and locally designed solutions.       

PDIA is far from being a one-size fits all solutions once there is no preconceived solution to begin with. Start with inspiring “why” (purpose) and an ill-defined “what” (policy challenge) to figure out “how” (policy trajectory) along the road can be tough and – sometimes – a source of dizziness, disorientation, frustration and stress for technocrats. Policy implementers must learn to sub-optimal solutions and marginal improvement if they want to make peace with PDIA stop-and-goes. It’s always easier to go with the flow, play out activities safe (as depicted by those PERT workflows), delivery ineffective products, declare partial victory massaging data to fit in log-frame indicators and move forward to the next challenge without paying so much of attention on what was left behind. 

Maybe that is the best promise of PDIA: you will make real, meaningful progress and – sometimes – succeed. And if fail occurs, you will lick your wounds, do some reflection, brush yourself out and get back in your journey of transforming public sector. You will “fail forward”.  

Learn more about the Implementing Public Policy (IPP) Community of Practice and visit the course website to apply.

Leave a Reply