Ok, thought of a number 4...

4. Given 1-3, it's not in the interests of governments to promote research into predicting earthquakes

...especially if they can't be predicted with total accuracy into a time-frame of a few hours. To get to that level of accuracy, and I assume we will eventually - where earthquakes can be predicted down to the minute - there'll be countdown parties by reckless youths on football ovals - but to get there we naturally have to go through a lot of less perfect levels

Which is not considered desirable coz they will cost a lot of money in lost work-hours, evacuations, and transport disruption as well as cause public panic

That gives you two options - or a combination of the two:

1. Discourage promising lines of research into earthquake prediction

2. Suppress information on partial solutions from becoming public

The thinking could be:

Whether you can predict an earthquake or not, they are going to happen, and cause major destruction. Although not fore-warning will inevitably result in some loss of life, it will also save billions in terms of the disruption costs

Personally, I think the information should be released whether it causes panic or economic loss or not. It's a very slippery slope when you start repressing information that will knowingly result in the loss of life