Why might Case Shiller futures clear at a discount to expectations?

I've been including language in many of my recent blogs about my belief that longer-dated CME Case Shiller home price index contracts seem to clear at a discount to expectations.^1 I've touted that contract prices are where risk clears, and are not necessarily reflective of expectations. In addition, I've also had to reconcile those views with the reality that as contracts "cash-settle", prices and the index must converge. So, rather than repeat the arguments in blogs going forward, I'm going to list all of the points here, and then reference this blog for anyone wanting a more detailed explanation.

As an illustration, let's start with an example where there are three sets of users that each have different forecasts. I'll refer to the groups as the 2% HPA camp, the 3% camp, and the 4% camp. Given their views, each of their "anticipated" forward prices can be mapped our to a four years horizon.^2 If the groups are equally weighted then the consensus view will be the 3% path. As 3% HPA takes prices to 112 four years forward, one might expect the futures price across this community to clear at 112 over the next four years -absent any changes to expectations.^3 ^4

However, my view is that two themes seem to bias clearing levels lower:

1) Housing bulls have many ways to buy exposure to the sector. They can invest in a variety of assets, and since real estate tends to embrace the mantra "location, location, location" there seems to be preference toward asset selection versus sector allocation. Also, many real estate investors have an income-oriented approach/mandates (hence the popularity of REITS that gives income favorable tax treatment), so there seems to be an investment bias toward the income side, rather than the price side, of investments. Finally some international investors may be looking to invest (and preserve) capital in the relatively safe United States, and are not necessarily looking to take leveraged price exposures (and with entities requiring KYC disclosure). Net, while I can make the case that insurance companies, pension funds and endowments might consider CME futures for incremental exposure to home prices (possibly as a hedge against unexpected increases in inflation), I've had limited success to date.

In addition, most bulls (at least the ones I engage with) want to invest in spot assets that will participate in the price rises they expect. Until recently, this has been a foundation to the boom in co-investment programs (e.g. Unison, Point) where one might buy a portion of a price at today's levels (while receiving zero cash flow), primarily with the expectation that higher future prices would lead to leveraged price returns. By contrast (using the above graph as an example) if the consensus view is that home prices will rise 3% per year, there's no upside to buying the futures as the 3% HPA gains are already priced in to the 112 price.

On the other hand, real estate bears have fewer tools to express a negative (or even less bullish view). This is particularly true for homeowners (the most likely participants in a market with small trades), as the approach of selling your house and moving the family into a rental (to ride out the expected collapse in prices) is fraught with logistical challenges, potential family disputes, and expenses. It might be more likely to make such changes if the population of homes for rent and homes for sale was more similar, and located in the same areas, but that doesn't seem to be the case in many communities.

As such, I'd expect a higher percent of bears than bulls would use futures, leading to a less-bullish bias in the CME Case Shiller community (between natural longs and natural shorts).

Further, many home price hedgers view futures as a form of home price "insurance". Much like when they buy car, home and life insurance, they don't expect to enter contracts at the expected value, but realize that in selling at a price below consensus (or for insurance, paying more than the expected payoff) they are improving their utility as they've hedged a tail-risk life event. Further, insurance companies don't write insurance at expected value. They, like the speculators I discuss below, enter into insurance agreements because they see an attractive risk/reward payoff.

Net, Natural Shorts are more likely to use these futures than Natural Longs, and they may see benefits at trading below "fair value".

2) Since there have been an abundance of Natural Shorts (aka Hedgers), and fewer reasons for Natural Longs to participate, the balance has to be made up by Speculators. Note that the current limited amount of trading in CME contracts makes is especially hard for institutional longs to participate -even if they wanted. Risk managers (and regulators) of such organizations, have legitimate concerns about being too large a percent of open interest, and/or depth of these markets. They might be more willing to participate if volume were higher, but in a Catch-22, volume might not rise until they participate.

Speculators are more interested in a favorable risk/reward exposure than in buying at consensus prices. Referencing the first illustration, if they believe that there are 3 outcomes -the 2%, 3% and 4% HPA price paths, and they read that consensus view is 3%, then buying along 2% price path might offer an attractive risk/reward. As such, I might expect clearing prices in this hypothetical example, to be below that implied by expectations, tracking either some lower expectations path, the discount that insurance buyers are willing to concede, or a level that suggests an attractive risk/reward (while still converging to the index at expectations). I've highlighted what the 2% HPA expectations path would be ^5 to give a sense of the possible price discount.

Note also that as time passes, the difference between the 2% and 3% paths narrows, consistent with tighter bid/ask spreads on shorter expirations.

Finally, according to this logic, increasing volatility should lead to lower clearing levels. For example, if instead of the three camps being 2%, 3% and 4% HPA, they are 1%, 3% and 5% HPA, the consensus level will remain the same, but the 1% HPA path starts 4 points lower. ^6 That is, if speculators (the marginal buyers) follow that path, and they are only willing to continue investing with the same risk/reward, they will lower their clearing bids. Note that this applies even if consensus is unchanged. If volatility increases, and consensus drops (consistent with current markets), one would expect to see an even more dramatic drop in the lower HPA path, and the prices speculators are willing to pay.

My sense is that this is the reality that these markets currently confront. The solution to reducing any discount vs expectations that speculators might want, is to either bring in more natural longs, and/or to attract speculators to compete with each other.

As I've noted, the key challenge to the first is that many institutional natural longs would first like to see a deeper market. However, increasing the number of smaller natural longs is possible. My blog from this weekend https://www.homepricefutures.com/posts/the-basics-of-hedging-home-prices-from-the-long-side on how to hedge from the long side, is an explanation to smaller natural longs (e.g renters and those looking to buy in the near future) as to how they might hedge a future event. With clearing levels lower, they might be able to buy home price exposure at 2017 prices.

I'd be interested in getting feedback from users of these ideas. My goal would be to incorporate any comments into updates of this blog, and to have this blog highlighted on the Resources page https://www.homepricefutures.com/resources for future use.

As always, please feel free to contact me if you have any questions on this blog, or any aspect of using home price derivatives in hedging strategies.

Thanks, John

^1 Among other observations, quotes on longer expiration CME Case Shiller futures contracts always seem to be below results of the quarterly Pulsenomics home price survey of 100+ economists, even when adjusting for index differences (i.e. Zillow vs Case Shiller). Also most Wall Street research forecasts that reference Case Shiller, seem to be higher than CME forward prices.

^2 For ease of illustration, I've ignored compounding so the 2% HPA path has forward prices of 102, 104, 106 and 108.

^3 Again for simplicity/illustration, I'm assuming zero interest rates, and no cost of tying up capital.

^4 Also just a personal gripe here - of course if the 4% HPA community thought final values would be 116 and the 2% HPA community thought final values would be 108, one might expect the two groups to trade between them. In fact, while I'm showing relatively small differences in HPA forecasts, and final prices (i.e. 108 vs 116), the Pulsenomics survey participants include much more outlier forecasts. The 2020 Q1 survey had forecasts for gains through 2023 ranging from -7% to +43%. Alas, even with such differences, I'm not aware of any participants opting to trade housing futures.

^5 The graph blends what happens if a 3% HPA is realized combined with ongoing 2% gains. So, for example, after one year of 3% HPA, which take the then spot index to 103, an HPA of 2% each year going forward would take forward prices for the 2% camp to 109.

^6 A decline of HPA by 1%, (going from 2% to 1%) should translate into forward expectations (for that group) being 4% lower.