Good points all. Point (a) is a potentially a huge problem, I agree. An entire country suffering from poor pension investments all at once could have disastero

Good points all. Point (a) is a potentially a huge problem, I agree. An entire country suffering from poor pension investments all at once could have disasterous consequences. My understanding of the typical pension opt-in change is that it opts you in to your company's default pension rather than a state default pension, so the nation as a whole would have diversification even if customers of some pension providers are at risk (unless that nation didn't have diverse pension providers, in which case they already have a problem). Could be wrong, not my area. Issue (b) is interesting too. I'd say it fell into my third group because it seems to rest on people being lulled into a false sense of security and then taken advantage of. There are a couple of reasons why I don't think it'd be a problem... one applies to "good" nudge units and the other is more general. Firstly, at a 'good' nudge unit they operate transparently, both at the point of intervention and when publishing findings after the fact. If you look at Richard Thaler's work, he uses 3 'laws' to define 'good' nudges that make it hard for abuse to seem possible: http://www.nytimes.com/2015/11/01/upshot/the-power-of-nudges-for-good-and-bad.html http://www.nytimes.com/2015/11/01/upshot/the-power-of-nudges... Secondly, a more general point. Nudge units on the whole use a tiny subset of the psychological arsenal that companies use daily. They only use the stuff there's clear scientific evidence for ploughing on layers of gamification, incentives, information assymetry, etc. Yet companies do not feel bound by Richard Thalers "3 laws of nudging", they rarely follow more than 1. That's fair enough, as it's harder to switch to a different government and governments have more power over us, but it does raise the question... if our decision making skills atrophy over time then that would have already happened with companies, and I'm not sure I've seen evidence to support that view. Again, could be wrong.