[sokoloff]

It seems like it would also include people who are buying a new property to occupy as owners that will be mortgaged, while planning to later sell their existing, also-mortgaged, owner-occupied property.

Those people are not investors by any useful English-language definition, but would be counted as such by this method. This is exactly the situation we were in when we bought our current home, and anecdotally, is not uncommon when people are moving in together from previously separate, owned properties. Sellers prefer zero sales contingencies over one sale contingency and prefer one over two.

[__derek__]

I believe those would be called repeat buyers under the Bank of Canada methodology because the Transunion data should connect their new mortgage with the discharge of their previous mortgage.

[sokoloff]

If someone had a house with a mortgage, bought another house with a mortgage in May, and discharged the original mortgage in October (or the following February), were they a repeat buyer or an investor? Plain English would say "repeat buyer", but I can't tell definitively how they'd be treated in the data.

The one clear reference in the article strongly suggests they'd be categorized as an investor: "The central bank defines an investor as a buyer who took out a mortgage to buy the property while maintaining a mortgage on another home."

[__derek__]

I don't know. There's no link to the Transunion data set, so it's not possible to dig in. Even the sentence you quote is ambiguous because "maintaining" indicates an ongoing active state. Assuming that the dataset is a point-in-time snapshot, there are probably some repeat buyers classified as investors because of timing.