Embracing complexity does not mean rejecting simplicity in favour of messiness. Photo by John Carlisle on Unsplash

Global development is seeing an exciting paradigm shift. Increasingly, leaders and practitioners recognize that development is not a “complicated” challenge that can be neatly parsed out into separate problems and siloed departments, like assembling a car. Rather, the various tasks of development—poverty eradication, improving governance, climate action, gender equality, and so on—are all connected, making development a “complex” challenge.

The changes have gone beyond talk. Lead institutions—and in particular, UNDP—are now moving “complexity” to the heart of their mission. As one indication, the most recent UNDP Strategic Plan (2018-2021) uses the terms “complex” and “complexity” 16 times, compared to five times in the previous Strategic Plan (2014-2017). More importantly, the latest Strategic Plan concretely articulates the UNDP’s aspiration to transform its role, from a solution-provider to a solution-enabler and “integrator.”

While the turn towards complexity is extremely promising—indeed potentially revolutionary—certain pitfalls must be avoided. As complexity is new to development practitioners, there is little consensus, evidence, or even a basic demonstration case of what “embracing complexity” ought to look like in practice. Consequently, every organization believes it is embracing complexity by doing something different—when in fact, it may not be doing so.

Knowing “what not to do” is a useful first step toward determining “what to do.” Hence, at the UNDP Innovation Conversations in March 2018, I shared my research on “Complexity and Development 2.0” by highlighting what I see as three common fallacies of embracing complexity in global development.

Fallacy #1: Embracing complexity means to reject simplicity and embrace messiness

Fallacy No. 1 is the misconception that complexity means to embrace messiness. This leads practitioners to instinctively reject the very notion of complexity and reform agendas associated with it. It is a fatal fallacy and hence the first that I’ll address. 

Let me briefly illustrate the distinction between “simplistic,” “messy,” and “complex” through my research on economic-and-institutional development in How China Escaped the Poverty Trap.

A perennial chicken-and-egg debate in development is this: Is it good governance that leads to economic growth or vice versa? For decades, the canonical theories have argued either to “fix governance first” or “stimulate growth first.” By reducing the complex process of economic-and-institutional into a single big arrow, existing theories obscure a crucial reality: in fact, causality goes in both directions. Simplistic analyses lead to false debates.

If a simplistic analysis errs by reducing too much, a messy description errs in the opposite manner: no reduction. If an analyst described every part and every movement observed in a given case, without any effort to distill patterns, it would be like mindless rambling. Needless to say, we wouldn’t learn much from messiness.

A “complex” analysis requires reduction, but not too much. In my book, I show that the process of economic-and-institutional development can be distilled into this co-evolutionary, three-step sequence: harness normatively weak institutions to build markets > emergence of markets stimulate strong institutions > strong institutions preserve markets. My analysis adds an arrow that captures the non-linear, mutually adaptive dynamics of development. But the picture I present is not a senseless mess.

My figure below sums up the key distinctions among what I’ve termed simplistic, messy, and complex types of analysis.


What are the practical implications? Embracing complexity does not mean that managers that should “live with the mess” and abandon the search for patterns and order. Rather, it means that we should not over-simplify and believe that replicating best practices will predictably turn poor countries into Denmark.

More importantly, a complex analysis that captures non-linear, mutually adaptive dynamics in a reduced form can yield surprising and actionable insights. My analysis finds that the first step out of the poverty trap is to harness normatively weak institutions, rather than best practices, to build markets. This opens up a practical agenda: how might organizations like UNDP support low-income countries to use normatively weak institutions to build markets?

Fallacy #2: Embracing complexity means to always give more freedom and make fewer rules

A second common fallacy of embracing complexity is believing that we ought to always give more freedom, cut rules, decentralize, and grant autonomy. Behind this assumption is an intuitive logic—given that rigid rules in hierarchies constrain innovation, the solution appears to be to remove those rules entirely.

But this intuitive logic is misleading. In fact, effective adaptation and decentralization always requires appropriate structures, guidance, and rules—in my book, I call this principle “directed improvisation.” Direction and improvisation are necessary complements. In the absence of direction, ground-level actors either do not know what to do, or their attempts at improvisation may be chaotic, messy, and lack collective coherence.

The key question is, what “direction” is necessary for effective improvisation? One type of direction is deciding and signalling the appropriate level of autonomy and room for experimentation across different tasks. The political system in reform-era China presents a rich illustration of “directed improvisation.” In China’s context, Beijing directs, while numerous local governments improvise solutions to local problems, thereby generating a diverse and locally tailored pattern of development.


Beijing does not give uniformly strict or loose commands, as I show in Chapter 3 of my book. Instead, depending on Beijing’s knowledge of a given policy domain, it issues commands in three varieties: red (cannot be done), black (can be done), and grey (may be done). Ambiguity in the third category empowers local actors to try things out, creating room for what I term “bounded experimentation.”


For UNDP, this insight underscores the practical necessity for leaders or designers to first consider in which areas autonomy should be granted or constrained, and to what degree. We should encourage experimentation, but not too much and not always.

Fallacy #3: Embracing complexity means to be pragmatic and make fewer changes

Many observers intuit that because development is a complex problem, packed with numerous challenges, the “pragmatic” response should be to target fewer changes. For example, Jomo & and Chowdhury (page 18) recommend: “A more modest incremental approach involving a few important but feasible reforms… may be more pragmatic and likely to succeed.” Likewise, Rodrik calls for the “targeting of reforms on the most binding constraints”(page 5).

On the surface, these prescriptions make intuitive sense: if overly ambitious “big bang” reforms don’t work, then the opposite—targeting fewer changes—must be the solution. But again, intuition can be misleading. In fact, the effective way of transforming systems is likely to be a third option: making incremental changes across a wide spectrum of connected domains.

Why is that? The rationale lies in the concept of “institutional complementarities.” An incremental change in one area is more likely to work if a connected area also changes incrementally. For example, if an organization reforms staff training but does not change the metrics by which staff members are evaluated, then employees will not be seriously motivated to pick up new skills.

The figure below contrasts the three reform strategies I have discussed in pictorial form. China is famously known for its incremental reforms (also known as “crossing the river by touching the stones”). But to be clear, China’s pragmatism is not that of “targeting fewer changes,” as frequently prescribed. Rather, it illustrates the third variety of incremental but wide reforms.


What does this mean for organizations that want to retool themselves to tackle the complex challenges of development? My third insight highlights the necessity for designers to address these practical issues: What domains of tasks or responsibilities are connected? If one reform is introduced, what other complementary changes need to be made with it? This is “systems thinking,” expressed in actionable, no-nonsense, non-fuzzy terms.


Adapting development organizations to embrace complexity has the potential of making a huge difference on the ground. Reform-era China demonstrates the possibilities of “directed improvisation.” The reformist leadership enabled improvisation and locally tailored problem-solving within a massive bureaucracy of 50 million employees.

The first step to embracing complexity, however, is recognizing and avoiding its fallacies. Embracing complexity is more than just ingesting a new menu of buzzwords. It is a practical agenda involving concrete analyses of problems, data collection, and organizational design. My three conclusions are summarized below.


About the author
Yuen Yuen Ang is Associate Professor of Political Science at the University of Michigan as well as a Faculty Associate at the Center for the Study of Complex Systems. She is the author of How China Escaped the Poverty Trap (2016). This article is based on her presentation at the UNDP Innovation Conversation series on 28 March 2018. The author retains the sole ownership of images used.

UNDP Around the world