All of the above pleaseI agree on the gold buying. If the goal is to punish military we should go all the way. 75% greater cost to buying units as well. You could even take it one step further and make it +75% cost to upgrade units. This will also keep the system more balanced in the end game, as humans rarely build new units in the later game but focus on improving existing units.
There's a misconception in how distress hierarchy works.
When I was looking at the worst case, it means 0% need covered in all needs. This is rarely going to be the case, as it is impossible to not have some food and production in any city, so distress is never going to be 0% covered. Even the other needs, will only be 0% for a while, in very recently settled ones, where there can be 0 science and 0 culture, but that will change soon, with the first monument and the first council. Then, the size of the city will start to be relevant. In other words, controlling city size should be enough for keeping unhappiness low in any city.
Then.
It's not that distress is the highest need. It is that having high distress 'hides' other need issues.
For example, if only illiteracy is covered by 50%, and the other needs are fulfilled, then a 25% of the city population is going to suffer by illiteracy. But if there is also poverty in the city, and poverty is covered by 50%, then, instead of having 25% people suffering poverty and 25% suffering illiteracy, we have 25% suffering poverty and 18.75% suffering illiteracy.
In a scenario where 50% of the city population would be unhappy, the Marslow hierarchy makes only 43.75% of the city population unhappy. There's a 6.25% of the people that doesn't want to protest for their lack of education, because they prefer to protest for their poverty.
So, the method of using the remaining happy population is actually a way to reduce total unhappiness when there are several unfulfilled needs. The hierarchy just says which needs are hidden by which. Although it is not as simple to understand as a direct percentage of the city population (which can be as high as we would like), it is the advantage of giving a greater (subjetive) importance to some needs, that can be used as a suggestion to players on how to address their happiness problems (solve this first, then that).
It would be simpler to use a fixed percentage of the population, but I think it will prove to be more useful the percentage of the remaining happy people.
If we have each need have a 25% fixed percentage, then curing any 1 need cures 25% of unhappiness. With the 50/50/50/50 model, curing any one need cures only 12.5% of unhappiness, for a total of 18.75% happy population, since 6.25% is happy no matter what. So the maximum unhappiness is less, but curing one need fixes FAR less unhappiness. This helps us avoid the issue of curing only one or two needs while ignoring the rest and swimming in happiness.
From a UI perspective, what we need to show is that reducing any need fixes unhappiness by the same amount (while subtly suggesting the order of importance, as tu_79 noted). This part shouldn't be very tricky.
The tricky part is somehow explaining that when you fully cure Distress with 8 Distress / 4 Poverty / 2 Boredom / 1 Illiteracy, now you have 0 Distress / 7 Poverty / 4 Boredom / 2 Illiteracy
Seeing those other numbers jump up is for sure gonna make people FEEL the system is unfair unless we have a good UI.