The Associated Press reported yesterday:

In 2008, Obama had a 34-point advantage over Republican Sen. John McCain among voters under age 30. He won about two-thirds of the vote in that age group.

But a new Harvard poll suggests the president may face a harder sales job with younger voters this time around. Obama led Romney by 12 points among those ages 18-24, according to the survey. Among those in the 25-29 age group, Obama held a 23-point advantage.

It’s not clear if this is a blip or if the trend will continue. But it does point to an inconvenient truth for President Obama and his supporters: Mitt Romney doesn’t have to win among key Democratic constituencies. He just has to run better than Sen. John McCain (R-Ariz.) did. That marginal improvement coupled with a robust turnout in the GOP base will be sufficient to run Obama out of office.

Democratic spinners declare: “Obama will win the majority of the Jewish voters!” True, but will he get 78 percent and win among those voters in large enough numbers, say, in the Philadelphia suburbs and the Florida Gold Coast to win in those respective states?

The quandary also applies to Hispanic voters. Romney is trailing among Hispanic voters by high double-digits. True enough (at least for now), but in critical swing states can he take enough of the Hispanic vote to deprive Obama of victory in his must-win states?

You can go through nearly every segment of Obama’s 2008 winning coalition (in parenthesis are his 2008 totals for each group): young voters (66% under 30 years old), minorities (95% of African Americans and 67% of Hispanics), single women (74%), college graduates (53%), the wealthy (49 % over $100,000), Jews (78%) and Catholics (54%). In virtually every case he will do worse among those groups, if for no other reason than the worsened economy. To make matters worse for the president, Romney doesn’t need to close the gap with these groups in states he was never going to win (e.g. California, New York, New Jersey). He only has to cut the margins in swing states that will determine the race.

So if Obama’s base is a tad less smitten and won’t be “making history” this time, his erosion among these groups coupled with losses among independents (he won 52 % last time) will be enough to spell defeat. Next time you see polling on the candidates’ standing with a pro-Obama constituent group (e.g. Catholics, Jews or young voters) don’t pay attention to the net difference between him and Romney;consider whether he is doing as well as he did in 2008. Moreover, consider if he is doing as well in states that matter most. If not, he’s in some trouble.