For a number of reasons – poor survey design, uncounted incarcerated males, inclusion of GEDs, inclusion of immigrants, etc. – surveys commonly used by economists like the CPS provide poor measures of the US high school graduation rate. Heckman and LaFontaine attempt to fix these mistakes by comparing the CPS, decennial Census records, and annual high school surveys to create an accurate measurement of the US high school graduate rate. The current rate is around 77%, lower than the peak of 81% around 1970. Graduating rates among males are lower, and among blacks are far lower. Indeed, low education rates continue into college, with only about 25% of the current cohort graduating from college, and less than half ever attending college of any kind. This divergence in educational outcomes is likely a large driver of stagnant wages since the early 1970s, since wages among college graduates and postgrads have continued to increase. The US has fallen behind many other developed countries in the past four decades in both HS and college graduation rates. To my mind, these numbers are a national disgrace, and such low rates of human capital are absolutely drivers of future high crime rates and lower economic growth.
http://ftp.iza.org/dp3216.pdf (WP – full version in ReStat May 2010)