If it's a constant offset like you say, and given academia publishes and industry often doesn't, that might suggest it's dependent on the rate of advancement in academic research. Not in this field so I may be wildly off.