I have a love/hate relationship with centralization.
It’s a tension I see all the time. In federations, what is the optimal level of shared service to be provided by the national office to gain efficiencies, while still maintaining local personality and autonomy? In community development, where does it make sense to amalgamate back-office functions while maintaining specificity of neighbourhood or cultural preference? In philanthropy, granting committees would rather receive a single, coordinated funding proposal than a dozen similar, disjointed ones.
Efficiencies and economies of scale — yes. And exclusion.
When I was working on my PhD thesis in rural maternity care, many moons ago, it was widely reported that rural women in my province had excellent access to most aspects of maternity care. It was true, when the data was aggregated a certain way. (“Can you get a C-section? Yes. Can you get care on a weekend? Yes. Can you access a midwife? Yes. An ob/gyn? Yes.) Scratch beneath the surface though, and you soon saw that although these services were all available somewhere, they were not all available in the same locations, and women had to choose in advance (read: at the wrong time) the location where they wanted to give birth — and therefore which mix of services they would get at that specific site. (How do I know, when I’m three months pregnant, if I’d rather have an epidural or a C-section?) Centralized data made access look great — detailed experiential understanding painted a different picture.
All of this is taking me back to my recent rants in favour of collective planning. I believe it holds the potential to be the best of both worlds — better coordination and less unnecessary duplication are possible, but so too are conversations about when duplication is necessary for improved access, and when centralization risks painting local nuances with too broad a brush.
And, when well-facilitated, it accelerates progress. Collaboration can seem slow, but at its best, it proceeds thoughtfully enough at the beginning to pave the way for remarkable (and more inclusive) results later.
