The meta politics around all this is quite interesting. The principal objection today is in the means of departure rather than in the departure itself. But are there US programs that terminate slowly that weren’t originally designed to auto-terminate? Essentially, if you want a managed departure, you need to preprogram it in US politics because the mechanisms at play here are a lack of state power driven by a large number of action-blockers (probably better described by the Abundance people).

Since we never programmed in a departure into USAID, it ended the same way that our Afghanistan adventure did: in a sudden scramble after a decision was made by a sufficiently powerful political force. In some sense, if you want to prioritize a managed exit you need to program it in and reduce blockers to it. This is not optimal for you if you want to prioritize duration-under-policy because the same mechanisms reduce that.

It’s less that drastic change is because “someone got fed up with something” and more that we represent the outcome of power with non-linear policy because of enduring anti-action. So the suddenness is built-in because significantly altering things is impossible. You’d ideally want something like a near linear response curve with perhaps some amount of static friction so that changes occur in outcome with corresponding changes in political realities. I think we have something like what happens when you rub two sticky things together: stick-slip friction. They stick and increasing force doesn’t do anything until suddenly everything moves dramatically.

Whether this is a natural constraint of the problems we face or an artificial constraint caused by our political shape is not clear to me. After all, many things are best done suddenly: you can’t slow-roll a Normandy. But I think many other things that occur like this are an artifact of how we’ve constructed things.

If Medicare subsidies reduced over time or were more mobile between 0% and 100% of what they were would they have not gone to zero? If USAID could have gone down 33% and up 3x between administrations would that have led to net better results? It’s not clear. Sometimes the stability-for-a-while followed by sudden collapse is actually preferable.

But I think overall, these outcomes are less characteristics of The One Bad Guy Who Ended It All and more characteristic of the underlying dynamic[0] of multi-participant democracy with powerful veto power. Whether veteran benefit fraud or Social Security or pensions will be next is one for us to think about but the more of a third rail something is, the more likelihood the change in it is catastrophic when it does indeed happen.

0: in the same sense that search in an unsorted array is O(n) worst case. It’s not anyone’s fault. The universe has given us this interaction.