Mia’s question about relisting is important. I would keep the original appearance date where you can identify the same property, even if the agent or wording changes. Otherwise a repeatedly withdrawn townhouse may look fresh and pull the average down.
I would assume property-level variation until completed sales support the other interpretation. Five-bedroom townhouses within that price range can differ sharply by neighbourhood, condition and layout. Forty-three days alone does not show whether buyers disappeared, sellers started too high, or...
So the practical order is: define the charge, divide by property type and school boundary, identify relistings and withdrawals, then compare recent completions and price-cut timing. Only after that would I test whether higher recurring costs correspond with longer marketing. Right now, condition...
Recent completed sales should carry more weight than the -1.0% movement in advertised prices. Check whether sold properties had lower charges, better condition or earlier reductions. Even a modest number of clean comparisons could be more informative than the combined sample.
I’d ask both agents for the recent completed sales they used, then make them explain every adjustment for condition and exact neighbourhood boundaries. The higher number is only persuasive if comparable homes actually closed near it. Also look at withdrawn listings, not just active ones...