Bryan Caplan, over at EconLog, cites an interesting discussion on the role of education and productivity, as featured in Arnold King and Nick Schulz recent book From Poverty to Prosperity. The authors interview William Lewis:

AK & NS: One finding that might surprise some people is that the education level of the labor force isn’t nearly as important for the overall economic performance of a nation as commonly thought… How did you reach that conclusion?

William Lewis: … [W]e got the first hint of this when we were studying Japan back in the early 1990s…  There were many disparaging comments made in the U.S., and maybe even stronger ones made abroad (especially in Japan), about how the U.S. labor force was getting what it deserved because it was lazy, uneducated, and maybe even dumb.  But of course, the Japanese – the capable, competent Japanese manufacturing companies – showed that this notion was wrong by coming here, building their own factories, managing American labor, and taking a lot of other local inputs and coming within five percent of reproducing their home country productivity.
[…]
The great bulk of the evidence about education came from competent multinational corporations of any nationality, who showed that they could go virtually anywhere in the world and take the local workforce and train it to come close to home country productivity.  The clinching evidence [came when] we looked at some other industries.  We compared the construction industry in the U.S. with construction in Brazil and found that in Houston, the U.S. industry was using Mexican agriculture workers who were illiterate and didn’t speak English.  They were not any different than the agricultural workers who were building similar high rises in Sao Paolo, say.  And yet they were working at four times the productivity.
…Uneducated people can be trained on the job to accomplish quite high skill levels and quite high levels of productivity.  And that’s good news, because if the World Bank and everybody else had to wait until we revamped the educational institutions of all of the poor countries and then put a cohort or two of workers through it, we are talking about another fifty years before anything happens.  That’s not acceptable and it’s not necessary, thank God.