expectations – HomePriceFutures.com

Diagram to reflect changes in forward SFR markets

The graph below may be the best way of illustrating several points that are key to understanding home price futures, as well as the current debate over home price expectations:

All home price indices (including Case Shiller) measure events that have taken place a couple of months ago. Since traders often incorporate changing expectations, futures prices may go down (or up) while index values are rising (or falling). As the graph shows, the SFR (San Fran) index slowly increased in all but one month since Nov 2017, while the SFR X20 (Nov 2020 expiration) and X22 (Nov 2022) contracts rallied until about Sept 1, and then collapsed. As such, there may be useful, more-timely information in such forward-looking markets, that is not captured in published indices.
Futures prices are public, are updated in real-time, and are posted by traders willing to financially back their views. As such, they differ from periodic forecasts. Both may be useful (and more so, if markets are deep and liquid). The press and analysts, should pay more attention to futures prices.
Futures prices have to converge to index levels as the contracts settle on some future (albeit unknown) index value. The graph of the SFRX18 contract and the CS SFR illustrate this convergence. Recall that the data for the Nov ’18 settlement includes home price activity from July, August and September. As such, (particularly on such shorter-expiration contracts) expectations of future index values should dominate pricing of futures contracts, as subsequent events (e.g. stock market falling 200 points in October, can’t impact past index calculations).
At times, longer-expiration contracts may be more volatile than suggested by standard risk metrics. (See August-Sept 2018 section in graph.) This has implications for pricing longer-dated options.
The difference between curves may be consistent with views on forward HPI. The ~20 point difference between the X18 and X22 contracts in mid-year may have been consistent with ~3% HPA assumptions over the next four years. By contrast, today’s X20/X22 prices have fallen due to changes in views that may be more consistent with <1% HPA. A key debate in these markets is whether the forward premium collapse is due to more negative forward sentiment, or just the old “more sellers than buyers” adage. (BTW -The curves may go inverted -i.e. with longer-dated contracts trading at a discount to either spot or front contracts – as was the case in 2008. Such pricing would be consistent with negative HPA.)
Hedging works. Note that someone in May -July, looking to lock in market-implied HPA of ~3%, could have hedged spot SFR exposure with longer-dated SFR contracts, and made money on both sides as longs rose in value, while shorts fell, as HPA expectations collapsed.

Now I can’t make the above statements without adding some qualifiers and observations, that also speak to understanding trading these futures, and how one might interpret prices.

These markets are thinly traded. In the three contracts shown there may be have been weeks without a trade. “Closes” can change without a trade as they are (at CME) a function of last trade, a lower offer, or a better bid.
Traders may buy or sell (or post bids and offers) for a variety of reason beyond their expectations of future index values. For example, traders may be looking to hedge real estate (or other) exposures, or sense that sentiment will grow more negative. Given the lack of depth, the arrival of larger new contract buyers or sellers, may put upward/downward pressure on prices, unrelated to expectations.
I tend to show 1×1 “actionable” quotes (one bid versus one ask) to form graphs, narrow bid/ask debate. I may be aware of more interest, at, or inside bid and offer, but may not show (or keep live all day/week). As such: 1) please refrain from market orders that are bigger than amount shown, and 2) feel free to contact me if you’d like to share ideas.
The CME Case Shiller futures contracts seem to react most notably to infrequent events -e.g. monthly updates to index values, announcements on home sales, S&P 500 index changes of 50+ points. More trading tends to take place on such days, but otherwise prices tend to be quiet and trend/drift in-between with limited trades.
Most trading seems to take place in the first 30 minutes and last hour of the trading day (the former often coinciding with economic updates). A trader looking to get something done should appreciate (IMHO) that there will be more eyeballs on the contracts at those hours.

Feel free to contact me (johnhdolan@homepricefutures.com) if you have any questions on this blog, or would care to discuss any aspect of hedging home prices.

Thanks, John

Basics: Approaching Aug expiration -review of theory

With the August 2016 contract expiring next week (trading stops 3 PM New York on Monday, valuation as of Tues Case Shiller #’s 9:15) it may be useful to recap the (quarterly) expiration process.

The table below has the Case Shiller index history for 3 months (June, July, and August release) for 2015 as well as the last two months (June and July) for 2016. Quotes (bids, offers and mids) for the Aug 2016 contract prices are shown in the blue area. (Note that bid/ask spreads average just under 1.0 point across the 11 contracts, which is typical of this phase of the expiration cycle.) Finally, mid-market quotes are compared to last month’s release and the Aug 2015 release for MOM (month-on-month) and YOY (year-on-year) percentage changes.

I use mid-market levels to illustrate percentage gains as, in theory, if traders “know” (or believe in their research) what the CS #’s will be on Tuesday morning, they can bid below that level, or offer above that level. As such, since each point that one can buy below the index or sell above is worth $250/point, it has been argued that the bid/ask range should be consistent with some range around the expected Case Shiller release (as traders attempt to capture differences between what they “expect” each index will be, and their bid or offer). If so, the market quotes might be useful for those drafting next week’s press releases as it appears that market quotes are consistent with DEN, MIA, SDG and SFR all posting >7.0% gains for the last 12 months, while NYM and WDC will continue to be the laggards. ( I’ll leave it to other to explain why that might make sense).

Of course this theory would work better if the markets were deeper. Most quotes are only 1×1 (one lot bid for and one lot offered) and many of the quotes are mine.

The “accuracy” of the expiring contract has wavered from months where all 11 index values were inside the bid/ask spread on the last day of trading to last quarter where there were multiple outliers (some of which were > 1 point).

I’m happy to facilitate anyone’s efforts to express a view on the next week’s index levels. Please feel free to contact me (johnhdolan@homepricefutures.com) if you like to propose a trade.

Finally, I’d note that with the expiration of the Aug ’16 contract the CME will open a Feb ’18 expiration. Since that contract expires on the value of the CS index released in February 2018 (the Dec 2017 index), I hope to propose a list of Feb ’17/Feb ’18 calendar spreads for those that want to debate HPA gains (?) for 2017.

“Gauging Confidence in US Housing” Pulsenomics, Scatter Diagram

I continue to be a fan of the work done at Pulsenomics regarding the use of surveys to gauge sentiment related to home prices (either among forecasters, or home owners). The Pulsenomics quarterly survey is (IMHO) the best survey of forward home price expectations by those in the business of making forecasts (to include, in the interest of full disclosure, myself).

Recently, Cityscape published a detailed report where Pulsenomics founder Terry Loebs describes both the methodology in calculating the Pulsenomics Housing Confidence Survey. It is recommended reading for anyone looking for views of homeowners (and renters) toward many questions, including homeowner confidence.

The report may have particular use for those following the CME Case Shiller futures as the results are tallied for the 20 regional components of the Case Shiller 20-city index. Since ten of those indices have contracts traded on the CME, one might be able to use the survey results in their trading.

I’ve taken a first cut at comparing survey results against CME markets here. The table to the right takes a portion of the results in Exhibit 2 (using only those cities in the CUS 10-city index). (I’m only using the homeowner confidence index, but there is larger set of results that includes renter confidence. Readers are encouraged to look at the entire report and see how use of other data might impact results.)

To the Pulsenomics results I’ve added the CME trading symbols, the gain in the Case Shiller index for the last 12 months, the spot Case Shiller index levels, the bid and offer for the Aug ’16 CME contracts, and the difference between the mid-market level of the Aug ’16 contract and spot levels. (I used the Aug ’16 contract to minimize seasonal levels. Ideally, I’d like to use a longer-dated contract, consistent with longer-dated expectations, but wanted to avoid seasonal issues. In addition, many of the Aug ’16 contract prices are mine, so caution is advised as I’m evaluating my own inputs. That said, anyone who wants to narrow the Aug ’16 bid/ask spread, or to challenge the bids and offers, is encouraged to post prices. Hopefully the following results might prompt some such reactions. Also, I use the Case Shiller indices as those are the references for the CME futures. Pulsenomics tends to focus on the Zillow indices.)

I then took the Pulsenomics Homeowner Confidence Level and contrasted it versus the mid-market implied gains for the Aug ’16 CME contracts to see if confidence levels are correlated with higher home price expectations.

The scatter diagram suggests that confidence levels seem to be generally correlated with one-year forward gains. For example, CHI has both the lowest confidence level as well as CME Aug ’16 prices with the lowest implied gains, while SFR has one of the highest confidence levels and highest CME price gains.

The outliers (e.g. DEN, WDC in particular) might be worth evaluating more thoroughly. DEN does have one of the stronger improvements in confidence levels over the last year, so momentum might play a role there, but WDC has both one of the weakest trailing 12-month gains in the Case Shiller index as well as only modest 6-mon and 1-year gains in confidence levels among homeowners (as well as weak gains among renters (not shown here)).

Might WDC CME Aug ’16 prices be too low, or might confidence levels be too high? I have no strong views BUT am very happy to accommodate anyone looking to debate the topic, and/or to trade WDC (or any other) futures.

Please read the entire Citiscape report and/or feel free to contact me (johnhdolan@homepricefutures.com) if you have any questions.

May month-end recap posted to Reports section

I’ve posted a month-end recap of issues related to trading in the Case Shiller (home price index) Futures to the Reports section (or you can link here). The report has many of the usual data, graphs and tables, on trading, open interest and prices, but as there was not a complete set of bids and offers at month-end, some of the graphs (from prior months) are missing.

The graph to the right tells most of the story for the last year. Prices (as measured by mid-markets) continue to move higher. While back-month price moves are consistent with a change from 3 to ~4% HPA, it is the move in the front contracts that has contributed to the recent upward shift in the curves. Aug ’13 (Q13) and Nov ’13 (X13) quotes reflect a further significant jump in index values (most likely based on the time of year (seasonal factors) and recent change in sentiment.) Expect Wall Street forecasters to raise their 2013 predictions if the Feb ’14 contract prices are reflective of expectations.

I would note that I’ve added a purple line to this graph that represents closes. I included it to show how, in a strongly rising, thinly traded market, closes are often just the best bid. (The same phenomenon held true as the markets sold off in 2008-09 where recent offers often defined closes.) As such, you may see bids inching higher to (among other things) bring closes more in line with mids.

Another casualty of trending markets is that bid/ask spreads tend to widen (or the offers disappear), and trading slips to near zero. That has recently been the case here.

I’ll have more to say on that, and other issues in the report in future blogs. For now, I just wanted to get this out ASAP.

Feel free to forward a link to this blog (or the Report) to others studying the contracts. As before, anyone is welcome to contact me (johnhdolan

The Turn

With the turn in most home price indices, CME Case-Shiller futures have turned higher -particularly in the front contracts.

The graph below shows the value of each of the 11 CS indices as a percent of the most recent spot index. Values to the left of July 2012 are the historical CS index indices, while values to the right are those for the mid-market of the CME futures. (The black square is the spot price divided by the spot price or 100% for each index).

Note that the turn began in earnest with the June CS release. The July index saw month-on-month increases of as large as 3.9% (for SFR) and the August contract values center around another ~2% increase (for CUS). Nov 2012 mid-market prices are 4-5% above spot.

Much of the “pop” in futures prices occurs in contracts that expire over the next 6 months. After that, both calendar spreads and mid-to-mid market prices suggest modest ongoing home price gains. As such, the rally in nearby contracts is consistent with many research reports that price rises are potentially a function limited inventory, a change in sentiment, and some investor money entering distressed markets. The back contract prices suggest that this price surge is due to such short-term factors and does not represent a return to 4-5% annual home price increases.

With the abrupt change in prices on the front end of the futures curves, bid-ask spreads have widened and trading volume has declined. (July saw 10 contracts traded versus 83 in May -the largest in a few years). I would expect that the CS indices need to settle down, and we need to get past Labor Day, before bid-ask spreads tighten up to prior levels.

One additional observation is that while the short-term price histories of the CS indices are quite varied (since Nov 2009), the futures contract prices seem to move in unison. This high degree of correlation seems to be at odds with the notion that different parts of the country will rebound at very different rates. I would suggest that the combination of wide outright bid-ask spreads, and the high degree of correlation in futures prices, makes this a ripe environment for inter-city trading. (After all a trader may disagree that CHI, for example, will be +10% (over spot) by Nov. 2013, but that trader may have very strong views on CHI vs the CUS index, or other cities.) I hope to blog on inter-city axes (and am willing to tout any inquiries I receive) in a later blog.