I've formerly written and debated at length about this topic on the Steam Forums, but since it still isn't acknowledged by the StarDock team (or, it would seem, the community at large) how broken the current economic model is I'm going to give it another shot here in hopes of this issue getting the needed attention.
First, let's clear a few things:
1) I am not inherently opposed to a system the penalizes an empire for growing very large - in fact, I think it is a really good idea to stack odds against a such empire and give the smaller ones a fighting chance, and it [attempts to] create the interesting choice between building 'tall' or 'wide'. It needs to be done right, however. Civilization V is an example of doing it right. FE:LH is an example of doing it catastrophically, game-breakingly wrong.
2) Even if the Prestige model from FE was perhaps less-than-optimal in certain ways, it is still superior to LH's model by miles. However, unrest-per-cities isn't hopelessly unsalvagable and won't have to be abandoned completely, but it will require a major rework.
With that out of the way, let's get to the question you are no doubt asking: what is the problem with the current unrest-per-cities model?
The current unrest-per-cities model fails in two major ways:
1) Unlike Prestige from FE, unrest-per-cities in LH does absolutely nothing to discourage early game city spam. The optimal way to play LH is to spam as many settlers as you can as early as you can and settle any location that is remotely safe from monsters. This is because when the number of cities is very low, so is the unrest penalty. Thus, any new city you found will only increase, never decrease, your gildar, population growth, research rate, mana income and - most particularly - your production. Many of you are probably familiar with the term 'ICS'... FE:LH's early game is 100% ICS, a model practically every other 4X on the market moved away from years ago for good reasons.
Again, this model stands in complete contrast to FE's Prestige system which meant settling any new cities would decrease growth rate in older ones, even with just two cities. While the old model was arguably overly restrictive (players should be allowed to expand more aggressively early to produce more interesting gameplay, managing a single city is watching paint dry), it still boggles the mind why the developers would move the game from a model that recognizes the problems with ICS and severely restricts expansion into one that completely encourages it.
ICS is bad for a 4X, but the much bigger problem with LH's model is:
2) Unrest-per-cities totally disrupts the flow of the game at a certain point. When you have 15 cities in your empire, any further expansion will produce absolutely crippling unrest: 15x3 = 45 percentage points of unrest, distributed amongst your most productive cities where it hurts the most. Adding a new size 1 city with its pitiful output is just not worth it at this point and will only harm your empire. Effectively, this means that beyond a certain point, the game completely prevents you from expanding. The early game is spent massively colonizing the land, but you then reach a point where it says "sorry, no go!" and become paralyzed. It is a total paradigm shift that just murders the natural game flow, going straight from unrestricted expansion into total denial of expansion in one swoop. When you reach this point, the gameplay completely slows down and you'll be forced to sit on your ass and do nothing for the next 50-100 turns. Why? Because the only way to lift the curse of massive per-city-unrest is to tediously wait for your Fortress class cities to level up! Fortresses, to make matters worse, are the slowest type of city to grow, even if you do everything in your power to support their growth, such as doing the senseless things the current model encourages such as building Fortresses in high grain, low production locations supported with both your Tower of Dominion, Palace and Consulates and create a city that will never be useful for anything other than unrest reduction. Consequently, no matter how smart you play, you'll be forced to sit and wait for your Fortresses to grow to lvl 5 for literally hundreds of turns (it's that paint drying thing again... talk about tedious gameplay!). During the full length of this time you'll be unable to attack your neighbours or do anything else of interest, even if you could easily crush your opponents, because doing it will in fact weaken you. You have to sit and wait.
The problem, as we can then see, is that the game's only real options for unrest reduction (lvl 4 Fortress: Prison & lvl 5 Fortress: Onyx Throne) are locked away in the buildings that it take the most time to reach, with logical necessity creating a disruption of the flow of gameplay as you wait for them to unlock. Solutions to the problem could therefore be:
1) Dismantling the whole broken LH economic model and returning to the superior Prestige model... Which we know won't happen at this point, although the game would be better for it.
2) A ceiling on the amount of unrest-per-city penalty you can get. Beyond 10 cities, any more don't addtionally increase unrest. This has the weakness of still making large empires much more powerful than smaller ones, a speed bump instead of a blockade. Then again, this game is about dominating the world of Elemental, yes?
3) More unrest-reducing improvements for cities. Halving the cost of the 'Bless City' spell in v1.3 works a bit like one and has reduced the issue somewhat by allowing one to better convert mana into unrest reduction, but it wasn't enough - especially since not every city will have a slot for essence, and it remains an extremely costly spell. A real new building(s) is needed - City Hall? It's fine to let these new buildings be expensive - even put a gildar upkeep on them - they just have to be there.
Many other solutions could be thought up, but something needs to happen. Current gameplay is: expand like mad, wait 100+ turns doing nothing, expand - and it is not good.
Thanks for reading!