My analysis of correlations between the Case Shiller 10-city index (and the Freddie Mac National Index) and the top 100 Freddie Mac metro home price prompted three simple observations (and further questions to address related to hedging strategies):
1) Users looking to hedge might be more inclined to use Ratio Agreements in metros with low correlations (as those with high correlations can have effective hedges with CS 10-city index)^1
2) Caps and Floors on RAs (defined as percent of mid-market prices) should be wider in metros with low correlations.
3) Metros can be highly correlated with a National Index, until they're not. (See move in Boise correlation -orange line - from 2013-19 70%, to 2013-22 48%).
The graph below is a scatter diagram of the correlation of the YOY% change in the metro index vs the Case Shiller 10-city index (and is about the same using the Freddie Mac National index), vs the slope of the ratio between the two. In effect, the diagram shows how correlated the metros are with the national index, and what is the beta between a change in the national index and the metro. Recall that this was generally a period of rising home prices so those metros with slopes >1 (e.g. Riverside, Tampa) tended to have higher price gains than the Case Shiller 10-city index, while those with lower slopes (e.g. Rochester, Chicago) tended to underperform.
As noted on the right side of the diagram, cities such as Atlanta, Fresno and San Diego generally moved in lockstep with the Case Shiller 10-city index over the last years, and tended to move (up or down) about the same magnitude as the national index (i.e. slope ~1.0). That is, in hindsight, users could have hedged these metro exposures using the Case Shiller 10-city index futures at about a 1:1 hedge ratio. Chicago and Grand Rapids moved in sync with the Case Shiller 10-city index but only moved with about 60% of the moves in the benchmark, so a smaller percent of hedges could have been used. On the other hand, Riverside was also highly correlated but was about 50% more volatile than the 10-city index.
At the other extreme (to the left side) metros such as Boise, Spokane, Syracuse and McAllen moved much more randomly against the Case Shiller 10-city index since 2013. For example, there were periods when San Jose outperformed and underperformed the CS 10-city index while CS was rising. Metros toward the left side of the diagram (e.g. Spokane, Syracuse and Rochester) would (in hindsight) have been better hedged by addressing the basis risk between each metro index and the Case Shiller 10-city index.
That is, the metros to the lower left would be better candidates for Ratio Agreements.
However, a key assumption to any such review of history is that the future remains the same. However, seeing that the correlation of Boise vs the Case Shlller index dropped from >70% (covering 2013-2019) to ~48% (from 2013 to 2022) is a reminder that assumed correlations (and bucketing here into better/worse candidates for Ratio Agreements) can quickly change.
Net, a hedger worried that the historical correlation between Atlanta and the Case Shiller 10-city index might become undone, might be able to rationalize the cost of entering a Ratio Agreement (or any other hedging strategy that addresses this basis risk).
I sense that I've only seen the edges of what a diagram like this tells. Please feel free to steer me toward any further work in this area that might help users with their hedging strategies.
Thanks, John
^1 -I've run the same table using the Freddie Mac National Index as the comparison to each metro, and the correlations remain about the same. (No surprise as the Freddie Mac National index was highly correlated with the Case Shiller 10-city index for 2013-2022.