The Clinton administration has proposed a host of programs to create more jobs, better jobs, and higher-paving jobs. Yet there is no evidence that his program of "job training" and "education" will work. The administration, like many economists, has uncritically accepted a common myth about jobs.

The most important way to improve labor productivity is not "job training," but work experience. In 1991, male high school graduates who worked full time and year round, ages 18 to 24 years old, were paid an average of $16,559 a year. But of the same group, ages 45 to 49, the men earned an average of $32,336 per year.

The more experienced workers are roughly twice as productive as the relatively inexperienced ones, assuming workers are roughly paid according to productivity. For college graduates, the earnings growth associated with experience is still larger. The gains for women tend to be somewhat less, but are still substantial.

The data suggest that labor productivity for a typical male high school graduate rises about 3% per year, from the time he takes his first job to when his productivity peaks in his mid-forties.

Because the work experience of females is more likely to be interrupted, it is more difficult to judge productivity growth from the data. But I suspect it probably approaches the 3% figure for men. The data again suggest that job experience for females accounts for a large proportion of the earnings gains.

On-the-job training is by far the best form of training. The workplace has an incentive structure that leads people to improve their economic conditions through learning. Three percent per year is deliberate growth that comes from work and discipline. This is more than can be achieved from government "job training" programs.

Robert Reich has suggested we fund "job training" programs with an additional tax on business. That would only increase labor costs and lead employers to reduce hiring. Likewise, extending more generous unemployment insurance would raise the wage for being idle and reduce job growth.

The second greatest contributor to productivity may be education. I say "may," because employers do indeed pay college graduates a lot more than high-school graduates. But that may be because college graduates tend to be smarter, more motivated, and more mature.

Education may in fact be largely a screening device to find the more able members of the population. Nonetheless, to the affected individuals education has a payoff, indeed one that has grown over time.

The evidence is equally clear that spending more money on public education does not lead to greater learning. Eric Hanushek of the University of Rochester has surveyed the literature on several occasions. A majority of studies, he concludes, show little or no relationship between the amount of resources spent on public education and the amount learned.

My own research with colleagues reinforces the Hanushek conclusion. Spending on direct instruction can modestly increase learning, but a very large proportion of school budgets go for purposes that have nothing to do with instruction or are even negatively related to learning. Even if there were a high potential productivity payoff to education and training, under the current public school system it would make little difference.

Our country’s real job problem has not been faced honestly. Public policy has failed miserably in getting some people to take the critical first job and stick with it. This is particularly true of racial minorities.

Between the years 1890 and 1930, there were no major racial differences in unemployment. Yet according to conventional accounts, prejudicial conduct against racial minorities was more intense than today.

In 1954, the year of Brown V. Board of Education, 58% of non-white Americans of working age had jobs. In 1992, the pro-portion was under 56%. By contrast, for whites, the proportion working increased over time from about 55% to 62%. Forty years ago, then, nonwhites out-worked whites; today the reverse is true. This is despite civil rights laws that promise to correct the old imbalances, not make them worse.

Thirty years of federal programs designed to "help" poor people disproportionately affected minorities. They reduced work effort among the poor. Today, a black teenage girl with a baby considering taking welfare or a $6 hour job will usually take welfare. Welfare is worth as much as work income, yet she gets it without work. Government policy keeps poor blacks from taking the first step up the job-training ladder towards more productive employment.

How can we guarantee that everyone who wants a job can get one? The lowest unemployment rates in the 20th century occurred during the period 1900 to 1929, when federal intervention in labor markets was low. The single incident of double digit unemployment, the Depression of 1921, was caused by the effects of monetary intervention, that is, a rapid inflation followed by deflation.

The Clinton administration has proposed to do away with "welfare as we know it" by having government at various levels give jobs to current welfare recipients. That would indeed correct some of the incentive problems associated with welfare, but only if those new jobs pay high enough salaries. But high salaries paid by taxpayers can only add further burdens on public spending, leading to higher taxes, more inflation, or both.

A better solution to our current problem would be the one this country followed after the Second World War. In 12 months, the federal government reduced its employment by 10 million, the equivalent today, in relation to the labor force, of 22 million. At the same time, the government went from running a high deficit to running a substantial budget surplus.

The Keynesian economists of the era predicted massive unemployment. Instead, the annual unemployment rate never reached 4%. Markets handled the adjustment from war to peace beautifully, without job training programs, welfare programs, or civil rights laws.

Rather than adding new employees at various levels of government, we should start eliminating them. Recreating that post-war employment policy would be a fitting way to celebrate the end of the Cold War.