The second part of this is silly--no one at Citi is saying, "Hey, who cares how badly we do, Uncle Sam will always step in," as they see their bonuses getting chopped. The first part, though, is sensible to look at more carefully. But that also turns out to be wrong. In fact, I'd argue it misses one of the key points of the financial crisis.
One word that comes up in almost any discussion of risks to the financial system is "correlation," meaning how often events occur together. What we've seen in virtually every crisis is that bank failures and other economic catastrophes are highly correlated, in large part because financial players do not lock themselves up in rooms and gaze at crystal balls. They watch what everyone else is doing and then they do the same thing. This was true in the housing boom, this was true in the technology boom and subsequent market crash, and it was true, significantly, in the savings and loan crisis of the 1980s.
That last example is key, because in that case what happened is that many smaller institutions failed, but taken together they added up to a huge part of the industry. The problem of the giant institution that's an outlier and needs to be bailed out when everyone else is doing fine is one that exists only in theory. What happens in practice is that many banks, large and small, make the same mistakes and fail at the same time. In other words, it tends to be not single banks that need to be bailed out, but big swaths of the whole industry. Breaking up the biggest banks won't change that.
