When creating strategic rules, keep in mind that play will necessarily be unbalanced at either the strategic or tactical level. The whole point of strategy is to set up combat as a series of tactical encounters that are as unbalanced in your favor as possible.
Also, this all means different things to different people. When I've had this talk with b-school academics who study strategy for a living, they often disagree about what the word "strategy" even means. In wargames, for some players it will mean "very large tactical" or operational-level game. Others take it to mean that you play a general officer and worry more about supply, logistical train, administrative and traffic control issues-- but you're still counting individual mechs and moving them around hexes. For still others, you're playing a political leader determining objectives, assigning resources, and maintaining your own authority.
None of these options lead to the same game-- especially not when you try to mush them all together into one ubergame where you're juggling public opinion, staffing policy, and also deciding whether you'll fire both medium lasers or just one in this activation. Part of being a general means you stop trying to be a lieutenant, and part of being a Premier means that you stop trying to be a general. Historical examples abound of where someone tried to do just that and did spectacular damage as a result.
One approach to try is the "narrative strategic combat" style. Have an economic model, a wartime production model, and a political model for each faction. Each player, wearing his hat as Premier, sets economic and political policies or (acting as War Minister) works within a preexisting policy framework to achieve national objectives. Play at that level is balanced, and you're free to play as ruthlessly as you want to unbalance tactical situations for your troops to give yourself the biggest advantage. Game actions progress in "phases", an abstract measure of time that represents time before the strategic picture changes. One phase per month is pretty ok for busy adult players.
Each player has the option to declare one battle per phase to change some element of the strategic picture. This can represent either a typical battle or a turning point. Historians often use an encounter like 73 Easting to illustrate the progress of a larger conflict. And they love Historical Turning Points (hey, that's catchy), which are typically battles that mark a major shift in how things are going. Either the battle itself changes the outcome of the war, or its unexpected-at-the-time outcome illustrates a larger shift. Examples include Gettysburg, Kursk, and Tukayyid.
The idea is while the larger strategic picture is filled with unbalanced encounters not suited to the table-top, there are moments where things hang in the balance. For example, take a hypothetical civil war in Ambigua. The scrappy survivors of the overthrown Republic are in full retreat, while the People's Revolution rushes in with copious foreign support to secure the capital and solidify their power. In this Ambiguous example, the People's player declares a battle for the capital while the Republican player wants a battle to save his retreating forces. The players (or controllers) have to set up two balanced tabletop scenarios, setting forces, objectives, and outcomes in proportion to the larger strategic balance. The player with the least forces in the battle at the strategic level gets to choose between more play balance on the tabletop and higher stakes in the overall strategic picture.
So the first battle is a city fight. If the Republican player wants higher stakes, he can accept a very lopsided battle between his few remaining Mechs and the People's Revolution's most elite Mechs, backed by copious vehicles and infantry. His goal is to hold them for a set number of turns. If he wins, he keeps the city-- the overall force is holding out despite overwhelming odds. If he loses, the People's Revolution storms the city and the remaining defenders are annihilated. Let's say the Republican player chose instead for an even battle. In that case, it's a Mech vs Mech battle of both side's elites. That the People's Revolution will storm the city is given, so this represents a side battle between their advance guards as the Republic fights to preserve what forces it can in the retreat. Closer odds, lower stakes.
Meanwhile, for the other battle, the People's Revolution player has the disadvantage (not having enough forces in contact with the Republican force). He can choose between a conservative even-odds battle where he merely slows down the Republicans, or go heavens-to-Gretsky and risk confronting a larger but retreating force in the hopes of breaking their morale and causing mass desertions. Either way you shift the story somewhat, but men who desire great rewards must dare great risks.
This approach tries to self-balance. Some people hate playing in walk-overs, and so people in that situation can choose to play it safe on the tabletop with equal forces (still fun) and win at the strategic level. Players willing to be Bambi won't have a great chance against Godzilla, but if they do win, the strategic gains are sufficiently impressive to change the whole picture. Besides, people don't want to fight in the Battle of Who Cares Box Canyon. They want their victories and defeats to mean something. BT tries to do this (unsuccessfully IMO) by keeping the overall force sizes very low relative to the populations/economies/territories they're fighting over. I do this by declaring that the battles are historically important in advance, and why, and then deciding who happens to be there for the fight. That's counter to the merc approach of tracking a single unit through a campaign, but a sufficiently motivated storyteller could arrange for it to be your favorite Sisters of Mercenary unit that just happens to be at every key battle. Just like real Stackpole.
Notice how the bridge between the strategic and table-top levels is slathered in a thick dollop of intuition, BS, and handwavium. This is intended: that's why it's called narrative strategic. It requires that players be friendly and collaborative, but I really don't see why anyone ever plays in a group that isn't like that.