Background briefing for the Symposium on Institutional Innovation in Cities at the Bloomberg Center for Cities at Harvard University, 8/9 January 2026
By Geoff Mulgan and Caio Werneck, The Institutional Architecture Lab (TIAL)
This paper explores options for organising cities in the late 2020s and beyond, with a specific focus on cross-cutting tasks (from pandemics and emergencies to handling data or homelessness) and methods for organising horizontally. It looks at frameworks, cases from around the world and possible future design principles, with a particular focus on ‘mesh’ approaches including platforms, stacks, partnerships and other models.
<aside> <img src="/icons/glasses_purple.svg" alt="/icons/glasses_purple.svg" width="40px" />
Index
</aside>
‘Only a failure of imagination, the same one that leads people to suppose that everything has already been invented, leads us to believe that all of the relevant institutions have been designed and that all of the policy levers have been found.' Economics Nobel Prize winner Paul Romer
City governments across the world usually organise their work through functional hierarchies – departments or secretariats with specialised responsibility for transport, housing, sanitation, education, and so on. Their approaches mirror those of national governments and the traditional multi-divisional business which had separate teams for manufacturing, marketing, sales, and for different product lines.
Those hierarchical structures became the norm in the late 19th century and they still work well for stable, bounded problems. They ensure clear accountability; a concentration of specialised knowledge; and a means to engage relevant stakeholders. Often, they bring together officials and professionals with a strong shared ethos – whether for policing or education, transport or housing. But vertical silos have also always created problems. Many priorities don’t fit them neatly. Sometimes departments clash, or dump costs onto each other. They may fail to share vital information.
There is a long history of attempts to create more coherent, coordinated ways of working, and as cities face overlapping emergencies (from pandemics to climate disasters), and slow-burning crises (in jobs, care, security and housing) that cut across these silos, many are looking for new ways to coordinate action.
Some of the new options make the most of digital technologies which make it much easier to organise horizontally – with shared platforms, data or knowledge, or one-stop shops or portals for citizens. Some involve new roles (for digital, heat or resilience), new types of team or task force (such as I-Teams for innovation). And many involve new kinds of partnership or collaboration, with mesh-like structures instead of the traditional pyramid hierarchies of public administration.
This dossier is part of a broader exercise on how to imagine and build institutions for city governments across the world conducted by TIAL with the support of the Bloomberg Philanthropies. Here we combine examination of interesting cases from around the world with frameworks for understanding the challenges of coordination and using new tools to solve them.
Part of the aim is to liberate cities from the constraints of traditional 19th and 20th century pyramids, which still predominate in their everyday work. These administrative models still have their place. But in an era which has seen extraordinary organisational innovation (from TikTok and Google to Nvidia and Wikipedia) as well as the rise of complex new needs and demands, cities which rely too heavily on old models risk being neither sufficiently agile nor sufficiently trusted to thrive.
Cities have always faced cross-cutting challenges – how to grow, how to handle recessions or big influxes of migrants and refugees or at the extreme the challenges of natural disasters and wars. But today’s challenges – from climate emergencies and pandemics to problems of social cohesion – may be becoming the typical problems not the unusual ones, leaving a gap between the capabilities Mayors have at their disposal and the problems that are highest on their agendas. Most cities have organograms like this one from New York, still divided into functional verticals not so different from 50 or 100 years ago:
One answer to cross-cutting issues is centralisation – all the choices can go up to the Mayor and the central team, who are able to take a wide view. This may work for a few issues and at some times. But it is rarely viable for long because the decision-makers at the center are likely to have less good intelligence, including both formal and tacit knowledge, and because they risk becoming a bottleneck, slowing things down, simply because of the number and complexity of decisions to be made.
This is why any system needs a division of labor, a way to divide up roles and tasks. A Mayor can't do everything, everywhere, all at once. The key question, then, is how that division of labor is organised, and, in particular, how both vertical and horizontal roles are divided up. For tasks where the city administration has both power and knowledge – a clear view of what needs to be done and confident knowledge about how to do it – traditional hierarchies may be the most efficient option. But this is less the case for more complicated or complex problems which usually bring with them limitations of power (with power divided between tiers of government, business, civil society and so on) and limitations of knowledge (with uncertainty about what will work and why).
Organised crime; air quality; inward investment; housing; decarbonisation: all are examples where both power and knowledge are likely to be uneven and spread across different agencies. These therefore require different approaches: tapping into wider networks of intelligence to understand the problems, or get feedback on actions; and working more through influence, orchestration, weaving and coalition building rather than linear accountability.