U.S. Bureau of Labor Statistics (BLS) projections of occupational employment growth have systematically underpredicted the growth of occupations that require the most education and training. The latest data on occupational growth rates show that the BLS's recent projections of occupational employment growth to the year 2000 probably suffer from the same bias. Based on a regression analysis of trends in occupational shares, forecasts of occupational employment demand imply substantially faster growth of higher level occupations. A comparison of past and projected percentage rates of change in employment in high skill jobs to actual and projected rates of change in the stock of well-educated workers illustrates the supply/demand balance for college graduates. Findings show that, during the 1980s, employment in high skill occupations grew at nearly the same rate as the stock of workers with one or more years of college; employers wished to increase the proportion of workers in these ocupations who had a college education. A shortage developed and the wage premium for college graduates rose to unprecedented levels. Policy implications include the following: (1) the social returns to a college education are extremely high and likely to go higher; (2) continuing inequality in wage premiums will put U.S. corporations at a competitive disadvantage; and (3) public policy must focus on increasing the supply of technically and scientifically trained individuals. (40 endnotes) (YLB)