25 April 2013

No one gets to draw blueprints until they've layed bricks

Asymmetric Info | Megan McArdle | Obamacare Won't Be Doing Much for Small Business Next Year

I don't get the sense that at the time of passage, people had spent a lot of time thinking about the sheer mechanics of how this would all work: how the IT would be built, the rules written, the necessary information assembled. They spent a lot of time staring at the blueprints, not so much thinking about the building materials and the labor.
This is one of the advantages of studying CS: you actually take things from conception to execution. You can't just brainstorm up a bullet point that your system will have feature X or accomplish goal Y; you actually have to work out the rules for how to do X or Y and then implement those rules in such a way that X or Y actually happen — on time and without breaking everything else.

I don't think there are many other disciplines where that happens. I heard a story when I was touring Taliesin West that Frank Llyod Wright wouldn't take on a new student until they had actually built their own shelter on his grounds. We could use that kind of qualifier more often.

I think this is also a symptom of politicians failing to ask the crucial question and then what? You pass Law A, establish Department B, implement Regulation C. And then what? The vote isn't the end of the line. People change their behavior. The environment shifts. Incentives change. Agents react.

I want more lawmakers who are good at chess, poker, Civ, Supremacy... any game that forces you to think "if I do this, he'll probably respond by doing that, in which case I can do this, which will cause him to..." and so on down the line.

Attention lawmakers: you're playing an iterated game; act accordingly.


  1. I believe that the worst offenders of this sort of thing are the financial people, and the accounting/management consultants that they have spawned. In the financial world, it seems to be easy to just say “let there be money to spend”, with a click of the mouse, and then all sorts of things “happen”. As a result, they seem to developed the notion that the rest of the world works the same way. And, they have corrupted MANY senior managers in non-financial organizations who now have the same ideas about the way that the world should work.

    I first saw this disease when one of our senior managers got it into his head to hire some experts from Anderson Consulting (now Accenture) to help us re-invent ourselves and help us to work better. I thought that they would come in and do it the way a scientist or engineer would, by observing how things were done, talking to the people who actually did the work, and reading all of the existing documents. Instead they came in with pre-packaged PowerPoint presentations, talked only to upper management about what they wanted, and talked down to the staff about how poorly we operated, including lectures to the front-line managers that were degrading and demoralizing to all except those who drank the kool-aid.

    They tried to impose a commercial culture on a government regulatory agency that was/is constrained by a legal structure that has no parallel in business, and ended up grafting the worst aspects of business practices onto the standard govt bureaucracy. It was a disaster, and the manager who started it all was eventually moved somewhere else, after he made a decision that almost caused an accident.

    1. From what I've seen, some consultancies really do do things the way an engineer would (BCG founder Bruce Henderson was an aero engineer, for instance), but a lot just repackage the same bullet points like you describe. Similarly a lot understand the needs of their public sector clients, and a lot clearly don't. I suppose that makes it like most industries: some people get it, and others really don't.

      Robin Hanson made the argument a ways back that inexperienced consultants could still add value because the proper action was often pretty clear to everyone, but whichever stakeholder was going to lose would block reform. His claim was that bringing in even inexperienced outsiders can break the deadlock.

      I think the second point can make the first one worse. If you're being hired mainly because you can throw around the clout of the McKinsey or Bain or Whoever name then it doesn't matter as much whether you're the kind of consultant who gets it or not.

  2. I've been saying for a couple of years note that I don't think you should be allowed in government at any level unless you've spent some time working at a real job: cashier, construction, something like that. Unpaid internships of any kind specifically not included: you must spend, say, a year (to pick a starting point for discussion) working at a job where whether you get paid--and thus whether you eat--depends on your actual performance.