An enormous missing part of this seems to be that you are comparing the "return" of something that pays annually and a lump sum result at the end of the 20 year

An enormous missing part of this seems to be that you are comparing the "return" of something that pays annually and a lump sum result at the end of the 20 years. You'd need to compare it to an investment where you're also withdrawing the difference with the higher salary per year. Or, equivalently for this, have the extra money earned paid into investments each year. Really, part of what this is doing is comparing invested money to non-invested money. Edit - for a quick figure, let's assume you're getting the equivalent of a 4% annualised return on a $100k investment. That means you're earning $219k over the 20 years more, for simplicity I'm going to assume that's an even $11k each year. Our 7% example would be $507k over 24 years, invested all up front. Instead of $219k, we're actually at $480k. The two line up once you hit roughly 22 years in the workforce, at 23 you're slightly ahead. By 40 years of working, investing upfront is $400k behind, at a little over $2.3M vs ~$1.9M A way of phrasing the question to ignore working / etc, is which works out better $X invested at $Y as a lump sum. $Z invested each year at $Y. The question the article appears to ask is the better of $X invested at $Y as a lump sum. $Z kept as cash each year.