Gates commented that our schools are designed to meet the needs of fifty years ago: I think this is a good thumbnail description, and leads to the important question: why have they not changed? The key to finding a solution is to note that all organizations resist change, and the real question is how are some organizations successfully motivated to change.
Travelers to Russia still comment on how much Russia today reminds them of America in the past: during seventy years of communism, life changed relatively little for the vast majority of the population, and Tsarist elevators are still in use in some buildings. Stalin and Gorbachev, among others, forced through some changes that appeared radical at the time, but they could only do so by overriding the wishes of others (because the others largely wanted to prevent change) and so the overall rate of change was limited to what one man could envisage and enforce.
To get an organization to change, the individuals in the organization must be persuaded that change is possible, and that failure to change will involve more pain than the pain of making the change. Somehow, America was able to persuade its component organizations of this, but the Soviets were not, except when Stalin threatened to shoot those who failed to make a change he had decided upon.
I have frequently worked as a consultant at various 'Bell' companies and had access to their internal communications. While they enjoyed their monopoly status, management's pronouncements stressed their commitment to the wonderful ideal of ensuring universal access to phone service, and in general sounded like a watered-down version of Pravda. After the breakup, a very different message appeared. I remember in particular a speech in which a CEO said something like "To install a new residential phone line takes us eleven man-hours including paperwork. The cable company can install cable in six man-hours. As long as this discrepancy exists, the cable company will gain market share from us, and your job and mine are in jeopardy." Note that this addresses both criteria for producing change - six hours is possible as shown by the cable company, and failure to achieve six hours may be more painful than the required changes. In other words, competition is the only known way to ensure progress.
Now we see why the school system is not changing to meet the needs of the students and society - there is no competition between schools. Communications from the teachers unions in particular have exactly the same characteristics as those from the Soviet Union and pre-breakup AT&T: "we are the most virtuous people in the world, but we must do even better to achieve our holy calling". Clearly we must introduce real competition between schools, and a voucher system is the most obvious way of doing so.
Just as Stalin was able, at tremendous cost, to produce a small amount of change in the Soviet Union, Bill Gates' wealth can produce small local changes in a few schools. This is a long way from creating a permanent state in which schools will be motivated to change themselves.
As Worldcom showed, competition with inadequate oversight is dangerous. Immediately after an industry is deregulated, some leaders achieve great results by being the first to make particular beneficial changes. They become famous, and get treated like gods. As their competition catches up, they must try to introduce more changes to stay ahead. Since they started with the most obvious improvements, the new changes have less spectacular results, and some even prove to be mistakes. Since gods do not make mistakes, cover-ups are necessary, and because gods can do no wrong, subordinates go along with the cover-up. Similarly, in the school system today, the introduction of testing has prompted many schools to cheat in various ways. Before we deregulate the schools, a robust system equivalent to external audits is vital.
I once read a claim by the designers of the SAT or some similar test that the test was designed to predict the GPA of the student during the first year of college. That is right - a test should be judged by its ability to predict the future. The government should create a voucher system in which schools are paid according to the improvement in test results shown by the students. But what should the test try to predict?
The most obviously measurable part of the self-interest of the community in a student is surely that the student should grow up to become a taxpayer rather than a convict or welfare recipient, and that is what the test should predict.
When a student enters or leaves a school, he should be required to take external tests created by more than one company. The creators of the tests should be paid depending on how accurately they predict the taxable income (less welfare and convict costs) the student will ultimately produce. The school should be paid according to the increase or decrease in the prediction for each student between the score on entry and exit from the school.
This puts the testing company into a position equivalent to an auditor. They get paid more if their predictions are more accurate than those of the other testing companies. They, and only they, get to decide how to measure the factors that influence the future economic productivity of the student. The government pays them only for the accuracy of the prediction. The government also pays the schools extra when the predictions for a student rise while attending the school.
As the testing companies get better at predicting a student's performance, the incentive for the schools will change to encourage the characteristics that the testing companies have found to correlate with success. For example, a twelve-year-old daughter of a 28-year-old single welfare mother may be considered at risk for becoming a welfare mother herself, and the testing company will have an incentive to try to figure out what that risk is for a given girl as she enters middle school, and incorporate this in its prediction. The middle school then has an incentive to find ways of inducing the girl to exhibit characteristics that the testing companies have found to predict success for the girl, so that she will test higher when she leaves the school. Questions of whether this should be within the responsibility of the school are implicitly answered by this structure, and only effective action will be rewarded.
This structure is similar to the one that supplies us with household appliances. Manufacturers compete to produce appliances that meet the needs of the population, and the fire insurance companies set up the Underwriters Laboratories to make sure they are safe. The government has only to make the rules - the different competing organizations will then develop all the expertise needed to optimize the results.