I'm a little late getting back to it, but...continuing the musings on consequences in the sandbox...
Enough about the theoretical reasons why consequences are important. On to the practical aspects. How do I juggle multiple campaign world plot threads and determine and apply consequences? It seems to me that consequences for player action are a lot more straightforward than consequences of player inaction, so I'm going to focus on the latter here.
The first thing I like to do is to make a running list of the open adventure hooks I've presented to the players. How many should be active at a time? It's entirely subjective, but the optimal number of these, to me, is 3-4. I don't have any particularly scientific reasoning to support this; it just seems like enough to give the players some options without having to keep track of too much going on off-screen. Yes, there's a near-infinite number of potential adventure hooks in any given sandbox campaign setting. Only a few need to be "in the spotlight" at any given moment, though, and anything that you haven't drawn attention to or otherwise made known or necessary in the setting need not be tracked. Until it, or some effect of it, has been observed by the players, it effectively does not exist.
Probably the first hooks, the ones that start the campaign, will be dangled by the referee, but once the players get into the flow, they'll likely start pursuing their own goals and trying to unearth some of the "hidden" opportunities that better suit their own goals and tastes. Choosing which of the innumerable "quantum" hooks to shine the spotlight on becomes a collaborative endeavor between referee and players. Players reveal, through their actions and interactions with the places and creatures and NPCs of the setting, what they want to look for - legendary ruins to explore, injustices to set right, good deeds to do for the common folk, current political intrigues, quests for crown and church, or whatever - and the DM determines what's there for them to see.
Between sessions is the time to give some thought to how each of your open hooks is changing and developing. Depending on the nature of a particular hook, the situation may be very slow to change, or it may be constantly evolving. Either way, it doesn't have to be a huge struggle or eat up large amounts of your precious prep time to keep things moving forward.
Figuring out what happens to open hooks and plot threads is much more of an open-ended storytelling function than one of game mechanics, but if you'd rather not rely on pure DM fiat, there's a pretty simple way to inject a little randomness. Look at each hook, and decide how rapidly the situation is likely to change and evolve. For something like an old legend of a magical sword hidden deep in a labyrinth of ruins, the situation may be pretty stable, and the status quo may last years or decades. After all, it already has. On the other hand, if the bandits who have long been a petty nuisance get a new and ruthless leader driving them to audacious new levels of crime, things might develop very quickly.
Decide where on the scale each hook lies, and choose a die - a larger one like a d20 or d100 for glacially slow ones, and a small one like a d6 or d4 for those touch-and-go scenarios. On a roll of 1, the situation changes dramatically for the worse. A roll of 2 means a change for the worse, but not extreme or severe. Rolling the maximum possible result on the die signifies a dramatic improvement, perhaps even a resolution, while a roll of 1 less than maximum is a lesser improvement. Any result in between means that the status quo holds for another period. You might also make notes about the scenario's momentum, and add a +1 to the roll if things have changed for the better for two consecutive intervals (or a single dramatic improvement,) or -1 if they've gone downhill twice in a row (or a single dramatic deterioration.)
(I should point out that a d4 used in this fashion allows only for positive or negative change, no stable holding pattern possible, which might be appropriate to especially volatile situations.)
For example, let's take the legendary lost sword. It's been lost for generations, and everybody the PCs ask knows the story, so the status quo has held for as long as anyone can remember. It's not likely to change, but it still could. Using a d100 roll, a 100 might mean that a virtuous knight has succeeded in recovering the sword. A 99 might mean that some new information about the sword's whereabouts has surfaced. On a 2, a villain takes up the search. On a 1, he finds it. Anything else, and the sword remains lost and shrouded in mystery.
The bandit scenario is much more dynamic, so we'll use a d8. On a 1, the bandits might kidnap an entire caravan, including the wife and son of a local lord, whom they hold for ransom. On a 2, the gang attracts new members, and its numbers swell. A 7 might mean that local village militias enjoy some success, a setback for the gang's reign of terror. With an 8, a high-ranking member of the gang who knows of the leader's long-term plans could be captured. On a 3-6, the bandits continue their depredations, with no appreciable change in the balance of power.
Note that it isn't necessary to determine a list of specific results before rolling the die. The dice are simply a tool to suggest a direction, not to choose an outcome. I'd also suggest only rolling when there's some uncertainty to how events will or should progress. If there's a logical sequence that just makes sense, why muddy the waters with dice?
Most of the time, just writing down a sentence or two about how the scenario develops is sufficient. If there's a development that you're fairly confident your players will jump on as soon as they hear of it, abandoning their stated plans from the end of last session, write it up a little more. Provide news and rumors of recent developments during the session, if and when it's appropriate. Recap them at the end, when it's time for players to choose the direction for next session's adventure, so all the latest news that their characters are aware of is fresh in their minds.
Look for ways that open hooks and plot threads might affect one another. Does the villain who recovers the legendary sword go looking to challenge the bandit chieftain for control of the gang? Or was that villain actually the bandit chieftain in the first place?
Resolve and phase out hooks in which the players persistently show no interest, so you can cross them off your list. They don't have to turn out well - if you decide, with or without the guidance of dice, that the bandits succeed in taking control of an important trade route and charging exorbitant fees for safe passage, then let that become the new status quo until such time as the players decide to do something about it. Assume that the situation has reached some sort of viable long term equilibrium, and don't worry about updating it any more. Keep your number of "live" hooks manageable by retiring stale ones in a plausible manner whenever necessary.