The main cause of the Dems 1994 rout was structural. And most of the other causes, tended to play off or feed off that one, big reason.
Between the early 1970s and the early 1990s an entire region -- the South -- moved decisively from the Democratic to the Republican column. Something similar happened in the inter-Mountain West and in border state parts of the Midwest. But the full impact of the transformation was hidden by incumbency and the stretch of Republican presidential rule from 1980 to 1992. As long Southern Democrats tended constituencies and could selectively hedge positions and pivot off Republican presidents, most could hold on. But that made this leg of the Democratic majority extremely brittle...
The big game changer -- paradoxically, because he was a Southern Democrat -- was Bill Clinton... [T]he truth is that the pre-94 Democratic congressional majority was never going to survive another Democratic presidency. A Democrat in the White House, pursuing any substantial part of the agenda of the party who put him there, would deprive those members of Congress from the Greater South (South and overlapping border state areas) and West of that ability to balance and hedge. And so it did...
As you can see, if my theory is right, 2010 is fundamentally different. The key problem for Dems isn't unpopularity. It's a highly apathetic Democratic electorate facing an extremely energized Tea Party GOP.
Marshall is strangely agnostic in his introduction, however, about the odds of a GOP takeover of the Senate:
I wanted to address this question of what if any meaningful parallels there are between 1994 and the 2010 mid-term elections. The short answer is that I think the parallels are significantly overstated. That doesn't mean that the Dems couldn't lose one or both houses of Congress; they could. But if they do it will be for different reasons.
While it's of course technically true that Republicans could take back the Senate, the odds of them doing so are extremely low -- the Intrade futures market currently puts the probability at 6%. And as I've argued, a GOP House takeover, while more plausible, remains unlikely (the current Intrade probability is 35%).