Nick Manning’s two recent blogs (here  and here ) raise an important issue. On the one hand, people interested in development have big ambitions. We want not just more, but dramatically more people to be educated, healthy and prosperous, to name only three good things. If we are lucky enough to have some influence over governments and development agencies, we might be tempted to work from the top down to get what we want, turning those ambitions into public policies and programs, and rolling them out by the yard like so much cheap office carpet.
But on the other hand, the same human values that make us want those things make many of us sympathize with the bottom-up tradition that takes individual humans or small communities as its starting point. We know how a state planning juggernaut led to the terrible famines in the Soviet Union in the 30s and China in the late 50s. We know the horrors that followed Year Zero in Cambodia. Schumacher’s Small is Beautiful and James Scott’s Seeing Like A State are touchstone texts. Likewise, some of us have an instinctive preference for ‘searchers’ over ‘planners’ , ‘positive deviance’  and ‘problem-driven iterative adaptation’ .
Small is beautiful. But might it also be trivial? There have been so many development projects which delivered benefits to a few but could not be scaled up to benefit the many. They were just a flash in a pan.
So we are on the horns of a dilemma. On one side, the grandiose futility of top-down; on the other, the stunted ambitions of bottom-up. Is there a way of getting ourselves off the hook?
A good place to start is to look at the evidence of large-scale policy success, and what has produced it. The excellent ODI series on "development progress "; the variable but useful Princeton "Innovations for Successful Societies"  series; the World Bank’s "Yes, Africa Can " report on growth policies in Africa; Bebbington and McCourt’s 2007 Development Success  volume. They teach many lessons, including that success often has its roots in an upsurge of what Hirschman called ‘social energy’ , translated into a political imperative to increase growth, improve services, etc.; and that hostile stakeholder interests can be neutralized by smart policy framing and assembling a coalition behind the policy.
Development assistance has played a part in some of those successes (though not in many others). Nepal could not have saved so many of its mothers' lives without donor money and donor advice. But I am afraid that the studies also teach the negative lesson that “the traditional form of donor support – the large pre-programmed reform project – breaks several of the cardinal rules of effective reformism.” 
A second step (since this is a blog and I’m allowed to be informal) is to learn from cycling. In my own version of "small is beautiful," I am currently following the boutique Tour de France rather than the bloated steamroller that is soccer’s World Cup. Cycling’s most successful coach is Dave Brailsford, the leader of Team Sky which has produced the last two Tour winners, and also the boss of the UK national team which has won a slew of gold medals at the last three Olympic Games. In a recent interview , he revealed the secret of his teams’ success:
"We analyse the demands of the event we want to win and work back from there. We don't work forwards, we work backwards – it sounds obvious, but not a lot of people like to do it because it's tedious. It's boring."
What do we get when we apply that to governance and public sector management, or to institutional reform? We start with a problem that arises in a group reform-minded officials or citizens. It could be as small as school textbook provision in a single country (an example of Matt Andrews'), or it could be as big as the World Bank’s twin global goals of eliminating extreme poverty and sharing prosperity by 2030. Being problem-driven doesn't have to mean thinking small. In either case, we work backwards from the goal to the steps that will achieve the goal.
Easy to say, of course. “If wishes were horses, then beggars would ride” is what we say where I come from (Ireland). There will still be the iterative, incremental slog of having to drop one policy idea and pick up another, find policy champions ... sheer trial and error. We might need to institutionalize the approach. A "Delivery Unit" which focuses on our goals, such as the one that the World Bank has just set up to advance some of its priorities, would be one way of doing that, but there are others. Whatever: "problem-driven" doesn’t have to mean small, and having big ambitions doesn’t have to mean being rolled over by a Big State juggernaut.
Cycling backwards to policy victory: is it worth a try?