But Forbearance Can Be Agile, Too

As the common adage goes, it is always easier to get forgiveness than permission. In business and social settings, risk-taking can become its own reward, and the severest consequences for exceeding one’s authority are rarely invoked.

We should be careful, however, about unleashing this mindset across the incremental, deliberative functions of the regulatory state.

On the surface, the appeal of agile regulation is hard to dispute: Government will never improve unless it tries new things. Better to fail in forging bold frontiers than in avoiding them at every turn. And maybe steady-as-we-go leadership should yield at times to a kind of restless impatience. To many observers, this culture-shift represents exactly the kind of strategic disruption government needs most. Or so the story goes.

The agile playbook spurs action. It eschews passivity and ministerial rigidity. Its counsel is well known: Deploy project management schema wherever they might fit. Use holistic approaches, with cross-functional, un-siloed teams working together. Experiment with the unorthodox. Try new things. Make innovation transparent. Gauge improvement continuously, and then revise project specifications whenever it may boost energy and vision. Ultimately, test the bounds of business-as-usual, especially where agencies enjoy a measure of latitude over how they conduct their affairs.

Change is a fact of life, and enlightened stakeholders want government to be more proactive, anticipating how social conditions will evolve on the ground instead of constantly lagging behind the curve. A trustworthy government is one which never takes its own performance for granted.

Applications of these principles have become well known across a number of settings. One is procurement reform, especially in the area of software acquisition, where layers of agency review can be streamlined and competition among bidders enhanced. Another is digital service delivery in the form of open-data systems and processes of data governance within and across collaborating agencies. Reengineering dynamic processes of disaster response and resilience by using simulation modeling and resource management likewise provides a ready fit for agile project logic.

That logic also fit well with the kind of real-time adaptation and problem-solving that multilevel COVID-19 pandemic response showcased so prominently.

And yet, as exciting as these prospects may be, agile methods inherently lean towards doing something new rather than leaving things be. The risk is that innovation for its own sake may bias choice and make things worse overall. Proponents of agile in business settings have seen a “do something” culture sometimes morph into a kind of “do anything” atmosphere. Frequent “standups” and “scrums” pressure teams away from preservation-based lenses and towards change-modeling as a ground of being. The blurring of traditional expert roles and authority structures leaves fewer actors responsible for steadiness and conservatism, lest they appear too readily satisfied with the status quo.

Balancing these concerns means that agile methods should be applied circumspectly. We ought to temper the agile impulse, whenever gradualism and forbearance appear warranted. A number of related regulatory conditions appear relevant:

• Novelty. Neither rulemaking nor adjudication flourish where the pace of technological change and innovation are too rapid. The temptation to compare emerging situations to more familiar ones may obscure altered conditions. Analogizing cryptocurrency to traditional monetary forms, or online sports gambling apps to casino-based in-person sportsbooks, may shortchange the true nature of emerging transactions.
• Expertise. Objective expertise may not always be available when government needs it most. Firms have had machine learning and neural networks under development for several years now, but the 2022-2023 revolution in popular access to large-language models such as ChatGPT has caught public response rather flat-footed. Establishing a backbench of agency-based knowhow will take time to develop and, in the case of A.I., may remain beyond government’s grasp.
• Ripeness. Relative to a variety of new-economy products and services, optimal governance may not reveal itself until the stakes become more visible and comprehensible. In such circumstances, government’s response may appear lackadaisical when, instead, the regulatory process has found itself in a state of mindful repose. Phases of contemplation and deliberation are crucial to success, even though no agile enthusiast would recommend slowing things down.

Traditional policy analysis defines problems and opportunities carefully. Options for collective action are devised and then compared against operative decision-making criteria. In this process, it often makes sense to stylize existing conditions as a kind of baseline-state, one which may ultimately be preferable to untried alternatives at least for the time being.

Despite agile government’s preference for innovation, we should recognize that “let present trends continue” may sometimes be the best available choice. That may not sound nifty enough for agile, but patience and composure have a pretty good track record.

Larry A. Rosenthal is Senior Lecturer at the Goldman School of Public Policy, University of California, Berkeley.