There's a deal of nonsense written about the graduate premium, generally by those who display an utter misunderstanding of what it means. Such a one is Aditya Chakrabortty writing in the Guardian today; "a man who comes out of university with a BA in history or philosophy will earn an average of only 2.3% a year more than if he'd gone straight into the labour market" he blithers, seemingly unaware that the premium refers not to graduate starting salaries but lifetime earnings. Sigh.
OK. Imagine a time-series graph with income on the left side and age 18 to 80 on the bottom. The graduate line will start low and stay low in the 20s, rising in the 30s and continuing to rise gradually until the 60s when it finds a plateau for a while before falling. A craftsman line - say a carpenter - rises sharply in the 20s into the 30s, starts to plateau in the 30s then starts to fall from the 50s onwards. The areas between the two lines represent the net financial lifetime difference in income - with the graduate showing a substantial excess.
Any manual or craft trade will have an earnings pattern that mirrors the growth and decline of physical strength - a door fitter who can fit 8 doors a day on piecework in his late 20s will only be able to fit 4 in his 50s, unless it is a craft such as potter or silversmith that doesn't rely on physical stamina. All cerebral occupations will tend to reward experience and continue longer. The 23 year old chippie may wave his fat pay wodge at an equally-aged impecunious student but it won't last.
The remarkable thing about the graduate premium that has surprised economists is that it was easy to understand in the days when only 5% of us went to Uni, but less easy to understand now; graduates still earn a lifetime premium (though lower than before), though 50% of new job market entrants have been to Uni. Explanations?