Is the middle class really doomed?

I've now seen this video at several liberal blogs, and someone has to stop it. Apparently, that someone is me, since no one else has stepped up. Mine is a high and lonely destiny.

Warren has an intriguing thesis: that women going into the workforce has resulted in few real consumption gains to families with children because all the money is going to childcare, and to bidding up the price of houses in good school districts. Meanwhile, families are more fragile, vulnerable to outside events, because Mom no longer functions as an all-purpose backstop. Meanwhile, the government is not providing the things those families need: childcare, high quality education, a more generous safety net, health insurance. The result: more bankruptcies, less financial security. The talk is provocatively titled "The Coming Collapse of the Middle Class: Higher Risks, Lower Rewards, and a Shrinking Safety Net"

As you can imagine, this thesis is extremely beloved of liberals, who like its endorsement of more government benefits, while ignoring the fact that this could equally well argue for having women stay home.

Nonetheless, I think it's an interesting thesis, and having read the book, I find it eminently plausible. The only problem is that it does not actually seem to be true.

As a general matter, my problems with Warren's work are fourfold:

1. Her arguments tend to rest entirely on particular statistics; if you look at another statistic that describes the same thing a slightly different way, her results have a tendency to collapse. Indeed, she often seems to almost deliberately pick the most useless number for measuring the effect she wants to get at.

2. She often switches nearly at random from one way of describing something to another: from percentages to absolute amounts, from individuals to households. All of these switches have the effect of concealing the holes in her work.

3. She either isn't familiar with, or ignores, fairly standard alternate explanations for the statistics she uses.

4. When she creates her own measures, they use overbroad standards for the things she wants to measure, while leaving out important variables.

Why (almost) all of her arguments are wrong in particular is so long that I'll break it up into various posts. Warren is pretty much the public face of moral panic about credit, and as such has a lot of influence. And also, a lot of the things she says are common tropes that should be addressed, so this seems like a good opportunity. More in succeeding posts.