Government and the institutional memory of failure

I think that large, established organizations operate according to different incentives than start-ups. Large organizations focus on the downside of new initiatives. In the start-up world, the focus is on the upside.

Large, established organizations develop a “culture of no,” in which lots of people have veto power and there is little incentive to say yes. It makes sense, because a lot of people have a stake in the existing way of doing things, a major failure could cause an otherwise-thriving organization to get into trouble. So you can see why there was not much effort made by Xerox to dominate the personal computer market, even though they were the first to develop a lot of the critical technology.

With government, the culture of no gets even stronger. There is no competitive pressure to adapt or change. There is tremendous institutional memory of failure. The FDA can remember every drug that was approved prematurely and found to be dangerous. It has no memory of drugs that were delayed unnecessarily.

That was Arnold Kling in a much longer post about elites and institutions. Read the whole thing.

Ben Horowitz claimed that it isn’t about small or large organizations that make them entrepreneurial or bureaucratic. He said it is between young and old firms.

Facebook, Google, and Amazon, still run by their original founders, are considered “young” firms because their leadership remembers building something from scratch, taking risks, and understands that they can be upended if they don’t disrupt themselves.

Firms run by career managers who rose through the ranks, however, are old firms. Usually, there is a large number of approvals that need to happen for anything to move forward. It is in that layering that a culture of “no” becomes likely, because it just takes one “no” to stall or kill a promising idea.

There is no “founder” type experience to be had in government (though, perhaps DOD internal “software factories” like Kessel Run can be thought of that way). The activities of program offices are decided by the military staff well in advance of its assignment to a manager.

Defense officials usually care about whether a program is cancelled or not. There is a long institutional memory of programs that were cancelled like the FCS, DDG-1000, and the Comanche. These are used to justify killing an idea before it starts.

All performance metrics are about meeting staff expectations — about avoiding failure — rather than evaluating what performance was actually achieved for the cost. So long as the program isn’t cancelled, you rarely hear about performance degredation to the specification, even if it makes the system a liability. All you hear about is cost growth, which really doesn’t tell you anything about the worth and performance of the program.

Perhaps for government officials to look more towards the upside than downside it requires them to actually participate in the upside — whether than is financially, by promotion, or through awards and other honors. But for that to happen, program success needs to be identified with individuals, not with bureaucracies.

Perhaps all the workforce needs is the chance to contribute their creative energies to challenging and important problems like defense. Innovative types probably don’t appreciate executing marching orders, especially when they fundamentally disagree with them or believe that they will lead to substandard results.

Here’s another good part: