So we're not spending less per pupil because we're spending less on everything. We're not spending less per pupil because we have more pupils per capita. What the figure at the top left of the next page shows is that we're spending less per pupil, basically, because we devote less of our budget to it. The figure charts the combined expenditures of state and local governments, because—California's not unique, but we may be the extreme case—the admixture of state and local funds from program to program is so complicated that figuring out where one stops and another starts is frankly not worth it. In any case, I think you get a more accurate picture by looking at the combination of state and local spending. And, again, you see that we always have devoted, and continue to devote, less of the overall budget to public education than other states. In fact, if we pulled our spending percentage up to the national average, our expenditures per pupil would be a little higher than the national average.

Well, if we've got just as large a budget (adjusted for population, etc.) as the other states, and we're spending less on K-12 education, that means we've got to be spending more somewhere else. And, in fact, the one place where we now and for this entire period have spent substantially more per capita than the other states is law enforcement. (This is a slight misnomer; what I've tracked below is actually cops and prisons. Police and prison expenditures are about 75 percent of the total criminal-justice budget. Judicial administration and court expenses add another 25 percent or so to these numbers.) Right now we spend nearly 3 percent more of our budget—about 81/2 percent, compared to a bit less than 6 percent for other states—on cops and prisons. If we were to shift that money„about $900 per pupil, or, at 5.7 million pupils, roughly $5 billion per year—to education, we'd basically be up to the national average. Of course, that would mean having a lot fewer police cars on the street and a lot more bad guys walking around, but that's the sort of trade-off that policy makers must engage in. And in California we tend to err on the side of cops and prisons.

In 1968, the California prison population was about 16,000; now it's about 165,000. That doesn't include the very large batch of people in county slammers—to be counted as a prisoner, you have to be sentenced to a year or more. I think we're getting really good at building prisons, by the way. We can put a 4,500-unit, state-of-the-art facility on line for about 330 million bucks. We do the same thing with prisons that the French did with nuclear power plants—we just apply the same plan over and over, like a cookie cutter. There are certain economies there.

Is this bad public policy? The statistic I always read is that spending on prisons is crazy because it costs $30,000 a year to send a kid to Harvard, and about $75,000 to put him in Avenal. Well, a lot of people have done economic analyses of this, and prison actually looks like a pretty good investment if you're worried about the return to society. These studies say that inmates, were they not in prison, would each be doing some $100,000 per year's worth of economic damage to society on average. That's apart from what disutility you might get out of being hit over the head while you're being mugged. So on an economic basis, we could lock a lot more people up before the marginal cost of locking up the next inmate equals the marginal gain to society monetarily.