Case Shiller and Teranet HPIs: Follow Up 1

Following a previous post where I aligned house price index troughs to determine relative valuation, I put in a few caveats, one that Vancouver's trough (which I assumed occurred in 1998) may not have been its trough -- even in the 1990s Vancouver may not have touched its bottom and was riding the coattails of previous speculative excess. The issue faced here is that the data don't go back far enough to the previous trough in the mid 1980s. Nonetheless we can re-scale Vancouver's data set with its trough pulled back to 1990, the start of the Teranet HPI. Below are the two graphs under the two scenarios for Vancouver:

Wowzas, that makes a difference, but I reiterate the caveats here, that the Teranet HPI data don't extend back far enough to garner a complete picture if markets like Vancouver's had speculation in the 1990s. Likewise other cities, most notably Calgary, are likely understated due to lack of data in the mid-1990s. To flesh out this we would need to turn to historical non-HPI data to augment this data series or simply use a different dataset altogether.

Below are the trough-aligned graphs with Vancouver under two scenarios, including the nominal appreciation scenarios.

What do we do with these graphs? The biggest takeaway, for me, is to evaluate the conditions in Vancouver in the 1990s and determine why the city was on a secular growth trend since the 1980s, which is somewhat at odds with experiences in other cities graphed above. Were conditions in the late-1990s the true "baseline", a point relative to which we should gauge future growth, or should we look further back to the 1980s where prices were lower still. If the latter, valuations in Vancouver are disturbingly high.

I do not normally concentrate much on directly comparing cities' relative valuations for a host of reasons, not least demonstrated by the sensitivity of the results by slightly changing assumptions. I prefer to look at earnings and prices, the so-called price-rent and price-income ratios, when formulating valuation metrics. But here are the data graphed in all their glory.