President Obama's approval ratings are, as was probably inevitable in this type of economic climate, declining. The Democratic advantage in the generic congressional ballot is eroding (I see no reason to exclude Rasmussen from that average). Job ratings for congressional Democrats are also going down. Fewer Americans are self-identifying as Democrats, too.
At the same time, Republicans are not showing an increase in support. Fewer Americans are also self-identifying as Republicans, and the GOP has made up no ground in partisan self-identification. Republicans have not increased their numbers in the generic congressional ballot, even though they are closer to Democrats than before. The Republican Party is viewed as, or more, unfavorably than it was in late 2008. Further, congressional Republicans have not seen any increase in their job approval during 2009.
Overall, what we are seeing so far is not a shift toward Republicans from Democrats, but rather an increase in the number of people who dislike both parties and have become "undecided" as a result. As such, if 2010 was a presidential election year, I would say this environment was ripe for a Perot-style, third-party challenge to once again break into the double-digits of popular support. The best bet for such a challenge would be an anti-Wall Street General, given the extremely low popularity of Wall Street and the high favorability maintained by the military.
For such a challenge to reach 15%-20% national support, the American exceptionalist, Perot line of anti-trade, anti-immigrant, anti-war, and now, in our own era, pro-coal is probably the best bet. It wouldn't win, but it would temporarily shake a lot of voters loose. Such voters would mainly come from the Republican coalition.
However, 2010 is not a presidential election. As such, given the consistently poor performance of third-parties in congressional elections, it is highly unlikely that increasing dissatisfaction with both parties will lead to a third-party breakthrough in the midterms. Here are the national popular vote totals for all third parties, combined, in House elections since 1978 (more in the extended entry):
Third parties have not done well in congressional elections. Even the minor third-party boomlet of 1990-2004, which never produced more than two independents in the House (and Bernie Sanders was the only candidate elected on a third-party ticket), has waned in recent years. The reason for this is that the most prominent third "parties" tend to be little more than cults of personality for prominent individual candidates (Ross Perot, Jesse Ventura, Ralph Nader, Joe Lieberman). When you have to run candidates in 400+ congressional districts, the cult of personality model doesn't fare well.
So, we remain in a zero-sum, two-party game. While Republicans have not gained in popularity, in order to reduce the size of the Democratic majorities in Congress, they don't necessarily have to do so (at least in 2010). Simply closing the gap on Democrats would be enough. This could result in significantly reduced voter turnout, which would also benefit Republicans, as polling in Virginia currently shows.
Democrats are not going to be bailed out by a third-party, or by continuing dislike of Republicans. More than anything else, what they need right now is for the economy to turn around.