What are the most important lessons for Dominic Cummings?

I promised you an entire synthetic post of my own on that topic, but your own comments on that query were so good and interesting that I don’t feel I have much to add.  If you have not already, do read them here, recommended.

Comments

Now so proud of this wisely groomed and pruned part of MR.

Respond

Add Comment

One more, if I may: there should be something to be learned about his reform project from Stafford Beer's experience with the Allende government in Chile. Despite the very different motivating ideologies, Cummings' enthusiasm for using technology to improve information processing (e.g. through Dynamic Land-type visualization/manipulation setups) seems very reminiscent of Cybersyn.

Respond

Add Comment

The empire struck back and Dominic Cummings cannot actually hire any weirdos. And so Tyler's prospective synthetic post is lost to us, like tears in rain.

Nice! You also could've added a line about how this would've helped them on their five-year mission.

Respond

Add Comment

Respond

Add Comment

Disappointed not to get to hear Tyler's thoughts, but agree that the comments were very insightful!

Another article this week in the Economist concerns a policy experiment in Liberia that outsourced the primary schooling of 10 percent of students to eight private operators and measured the results (economist.com/middle-east-and-africa/2020/01/02/lessons-from-a-radical-education-experiment-in-liberia).

"Donors often say they want projects to be quick, cheap, rigorously evaluated and under the control of local politicians. Liberia’s scheme was all of that. It cost far less than a typical project (using just $23m of donor cash); it was set up in less than a year; and it was subjected to a RCT. It also survived a change of government."

Yet "gains were modest ... Pupils beginning in privately run schools could on average read 15 words per minute three years later, versus 11 in state-run classrooms ... (still behind the 45-60 words per minute deemed necessary to understand a simple passage and far behind the more than 100 words per minute that peers in rich countries can read)."

Here you have a seemingly exemplary policy trial designed to foster innovation through competition and allow for measurement -- the type of five-point prescription the Economist might give in its leaders -- but which produced decidedly underwhelming results.

The "lesson" I would think Dominic Cummings should heed is that public policy concerns hugely complex systems that cannot be precisely directed. The civil service may lack competence and innovation (though I think he exaggerates this: to a "maverick", anyone presenting resistance is an incompetent) but it is misguided to ascribe policy failures to that alone. It shouldn't surprise him if he sits down at the Big Data dashboard in his fabulously reformed civil service of misfits, presses the Go button and nothing much happens.

Gains were modest? Sounds like a 40% increase in reading speed in the private schools? Is that modest?

Of course your second point is true, that Governments are highly complex organisations that cannot be precisely directed. It is why they need to do less. Anything that can be done by the private sector should be done by the private sector for this very reason.

I half expected that as an objection! I don’t know that proportional comparison is legitimate. If they had risen performance from 1 wpm to 2 wpm would you say they had doubled performance? I think there is an expectation of convergence from such a low base. Also, the private schools had considerably better funding and the best performer had an increased dropout rate.

My point wasn’t that government administrations are hard to direct (though I agree) but that, even if you can direct government effectively, the challenges of implementing policy remain.

"If they had risen performance from 1 wpm to 2 wpm would you say they had doubled performance? " well yes I would! Of course it would be better to have converged on western performance, but without being racist, they may not have have the capability to do that. I believe most of the gains of good education are at the lower end. Smart people are very capable of dealing with a bad teacher as they can access material on their own. So this to me is actually quite a success. Of course as you say there may be other issues I don't know about that makes this less of a success. My view is building on your comment, if the private sector can do something just as well as the public sector, then it should be private, so that the public sector can focus on only those things that can only be done in the public sector. So the improvement you quote here would be, ceteris paribus, be a bonus on top of just matching the state performance.

But overall

Respond

Add Comment

Respond

Add Comment

Respond

Add Comment

Respond

Add Comment

Respond

Add Comment