A frequent question among our chattering classes is whether our children is learning.  The answer seems to be no, they isn't.  What would explain that?  George Will has the answer:

Moynihan also knew that schools cannot compensate for the disintegration of families and hence communities — the primary transmitters of social capital. No reform can enable schools to cope with the 36.9 percent of all children and 69.9 percent of black children today born out of wedlock, which means, among many other things, a continually renewed cohort of unruly adolescent males. 

If you think the solution–the only solution, the panacea, as it were–is a rise in teacher salaries then George Will is going to prove you wrong:

Chester Finn, a former Moynihan aide, notes in his splendid new memoir ("Troublemaker: A Personal History of School Reform Since Sputnik") that during the Depression-era job scarcity, high schools were used to keep students out of the job market, shunting many into nonacademic classes. By 1961, those classes had risen to 43 percent of all those taken by students. After 1962, when New York City signed the nation's first collective bargaining contract with teachers, teachers began changing from members of a respected profession into just another muscular faction fighting for more government money. Between 1975 and 1980 there were a thousand strikes involving a million teachers whose salaries rose as students' scores on standardized tests declined.

In 1964, SAT scores among college-bound students peaked. In 1965, the Elementary and Secondary Education Act (ESEA) codified confidence in the correlation between financial inputs and cognitive outputs in education. But in 1966, the Coleman report, the result of the largest social science project in history, reached a conclusion so "seismic" — Moynihan's description — that the government almost refused to publish it.

We've already established that teachers' salaries have nothing to do with output, haven't we?  But lo, continuing from above:

Released quietly on the Fourth of July weekend, the report concluded that the qualities of the families from which children come to school matter much more than money as predictors of schools' effectiveness. The crucial common denominator of problems of race and class — fractured families — would have to be faced.

But it wasn't. Instead, shopworn panaceas — larger teacher salaries, smaller class sizes — were pursued as colleges were reduced to offering remediation to freshmen.

Couldn't it be, however, that smaller class sizes and higher teacher salaries are goods to be pursued regardless of their effectiveness at fixing a social problem they're not supposed to be fixing?  Who could dispute that teachers ought to be well compensated for the very important work they do (I'll exclude myself from that work–what I do is not really work)?  What parent would not want her or his child in a smaller rather than a larger class?

More importantly, where is the social scientist who would claim that paying teachers more will remedy the various social problems produced–get this–as a result of income inequality?  Indeed, while we're at the correlation game, why don't we correlate family incomes and stability with the absence of well compensated, union labor?  Since Mr.Will is so interested in quantitative social science, perhaps he might find the results so alarming he'd refuse to read them until the Fourth of July, at night.

So to sum up.  Teachers' salaries may have nothing to do with educational outputs.  But that's not why teachers should have higher salaries in the first place.  Second, the social problems kids bring to school stem in no insignificant way from economic inequalities faced by their parents.  These may come together at school, no one expects the school to solve anything but what the school can solve.   But teachers and schools ought not to be punished just because they can't solve that which they aren't suited to solve.