My sense is that speculators on the long side of CME Case Shiller futures have a core bias in their favor. I offer my rationale, and two ways to trade the conclusion.
It's long been my theory that longer-dated CME home price futures clear at levels below expectations.^1 I note this for two reasons:
1) My observations on where contracts clear relative to contemporaneous forecasts, and
2) My sense that while the contracts might attract natural shorts (e.g. those looking to hedge), I've had a harder time getting natural longs to participate. That is, while I believe that pension plans might have an interest in upward surprises to home prices, I don't see them using these contracts, but do see billions of dollars going into specific assets. Net, that industry seems to value asset selection over upside to sector allocation.
That leaves speculators to offset the interest of the natural shorts. Theory would suggest that speculators don't necessarily want a 50:50 bet, but might be attracted to opportunities when they see clearing prices below expectations. The question then becomes -how does one measure expectations, to compare them against forward prices -and to see if speculators have a better than 50:50 bet.
While the quarterly Pulsenomics survey has been (and remains) a great historical source of expectations, ^2 results are based on the Zillow ZHVI . The differences between the Zillow and Case Shiller indices (i.e. which parts of country are covered, and how index is calculated) requires one to qualify any interpretation.
Fortunately Reuters has (somewhat) recently started a survey on home price expectations referencing the Case Shiller 20-city index.
The table below show that, during August 2019, they received 37 forecasts for home price expectations for year-end 2019 (and 36 for year-end 2020). ^3 Importantly the median forecasts for 2019 and 2020 were 3.0 and 3.2%
There is a huge overlap between the Case Shiller 10- and 20-city indices (as the 10-city components have about ~75% of the weighting of the 20-city index). That said, the CS-10 index references larger cities (e.g. NYM, LAX -typically those with higher home prices) while the CS-20 index includes the booming areas of Phoenix, Atlanta and Tampa.^4 Given the out-performance of lower-priced homes over the last few, the CS-20 has seen higher YOY gains than the CS-10. However, given the large overlap, the out-performance of the CS 20 (relative to the CS-10) has averaged about 40 basis points.
Adjusting the Reuters forecasts down by 40 basis points might then be consistent with expectations of gains in the CS-10 index of 2.60% for 2019 and 2.80% for 2020.
The final piece for comparison and to support a view on whether futures might clear below expectations is the prices on the CME Case Shiller futures.
The Feb '20 and Feb '21 CS-10 contracts settle on the index values through a period ending two months earlier. That is, the Feb '20 contract references the year-end 2019 index value and the Feb '21 references the 2020 index values. As such, one can compare the same index (expectations vs. futures prices) for the same time horizon (year-end 2019 and year-end 2020).
The mid-market levels for the two contracts are priced for a gain of 2.68% (the Feb '20 contract vs. the index value from Feb '19), and ~1.00% (Feb '22 mid-market vs. Feb '21 mid-market). While it might be no surprise (given convergence) that the 2.60% implied 2019 forecast from Reuters is very close to the 2.68% implied YOY index change in the mid-market of the Feb '20 contract, the 2.8% implied index gain from the Reuters survey is far above the 1% gain implied by the Feb '21 contract.
Net it appears that the mid-market of the CME Feb '21 (G21) contract is trading at a significant discount to implied expectations. The 1.8% implied difference is consistent with the contract being priced about 4 points below expectations.
The key questions then are: is this a mispricing, or a fundamental imbalance between buyers and sellers that might be attractive to speculative interest?
If the later, there are at least two ways to play this. one could buy G21 contracts on an outright basis, or buy the G21/G20 calendar spread (currently bid 2.0/ offered at 3.2/ vs. 2.3 mid/mids) looking for spread to expand over time.)
There's a number of assumptions in this blog, but the observation (conclusion) if fairly important, so please feel free to contact me if you have any questions.
Thanks, John
^1 Shorter-dated contracts will converge to index so, in theory, they should be trading on top of expectations as expiration approaches.
^2 I am a participant in the Pulsenomics survey
^3 I am a participant in the Reuters survey
^4 The CS-20 also includes Seattle and Portland which having been doing too well recently.