It's not really relying on the data, it's sort of more pointing out that markets do tend to follow each other (fully appreciate not many people will want to read full academic papers though.. generally the conclusion is the best place to start). It's probably less about prices from different segments impacting each other (correlation vs causation etc), more about the reasons why there was a movement in prices in the first place. There would be times when prices for units/detached move a lot together, other times less so.
So don't take it the wrong way and I agree with you to an extent... a moderate fall in oversupplied inner city units would likely have a pretty negligible effect on the rest (detached least of all). However, a change in one of those major drivers mentioned above (rates / inflation namely), then you're going to see all segments react fairly uniformly (micros through to high value detached, however detached obviously has a more valuable land component).
It's an extremely complex market, and at these low rates we should probably expect the unexpected. If inner city units were to collapse,purely because of supply, then it's hard to imagine detached is going to get taken down with it too much. But if borrowing costs for banks (rates in EU/JP/US) were to rise a fair bit, then that will most certainly flow through to all property values with oversupplied being hit the hardest (again there's the land component obviously for detached).
So.. summary: I don't see oversupply in inner CBD units affecting the wider market by itself immediately... but I do think prices move together to a certain extent (always in the same direction) across states, cities and type of dwelling over time.
