Despite their overall economic growth over the last few decades, countries across Asia still face the complex task of structural transformation. Low-income economies must build formal industrial and service sectors from agricultural and informal bases; middle-income economies must move up the value chain; and high-income economies must continually generate new capabilities at the frontier of innovation.

Meeting this challenge requires implementing and adjusting solutions addressing a range of problems--problems whose complexities imply often it cannot be known ex ante whether proposed solutions will succeed or fail. Agencies tasked with delivering rapid growth must be able to act both effectively and nimbly: trying potential solutions, discarding sub-optimal ones, and reallocating resources quickly.

The prerequisite for such "learning by doing" is flexibility, an institutional capacity with relevance to tasks beyond economic growth. It is a capacity easy to advocate but hard to build, facing difficult problems of exit costs (especially political), decision-making under uncertainty, and the governance and use of discretion within a bureaucracy.

This study focuses on how real-world policymakers might operationalize the capability to be flexible in the agencies they lead or create. It does so through detailed examinations of three types of organizations--venture capital, defense research, and industrial policy--chosen for their ability to flexibly deliver on a portfolio of investments or programs. It concludes by offering a tentative set of techniques and strategies to enable flexible implementation.