Zillow – HomePriceFutures.com

What Homeowners (and markets) are thinking

Pulsenomics recently published their semi-annual survey of homeowners. (You can order a complementary copy here ). The 18-page report is chock-full of surveys (using great graphics) on what people think about home ownership, willingness to buy, and (importantly for traders here) outlook on prices. The survey participants are divided into homeowners vs. renters, AARP members vs. Millennials, and many other ways. In short, I’d argue that anyone looking to trade CME Case Shiller home price index futures should allocate a good block of time to work through the findings.

The only thing that’s missing is an action plan should you agree/disagree with any of the home price forecasts for the next year. Fortunately CME calendar spreads may be the best publicly quoted market where you can express a different view.

Now, let me first caveat that the Pulsenomics survey references Zillow indices, while the CME settles on values from Case Shiller indices. Second, there may be many reasons why survey results might differ from market-implied HPA. That said, the CME contracts may be the best way to express an opinion on HPA for 2016.

The table above shows recent market quotes for the Feb ’17 and ’18 contracts. This is relevant to a discussion of 2016 HPA as both Feb contracts settle on the value of the Case Shiller index for the period ending in December. Thus, one can compare outright markets, or mid-market vs. mid-market prices to get a sense of percentage differences between the two contracts (that reference Dec ’16 and Dec ’17 values). (A caution -indices can be revised after the initial publication).

First, note that the mid-market of Feb ’18 contracts is higher for all 11 regions than for the Feb ’17 contract. Differences (expressed in percentage terms) range from +2.4% for WDC to +4.3% for DEN. Note also (in a point made in the Pulsenomics report) that the percentage gains are generally much lower than the most recent 12-month change (with the exception of NYM and WDC which are coming off under-performance vs. peers).

However, rather than looking at mid- to mid-market prices, a market (the calendar spread market) also exists where instead of selling (or buying) futures in one expiration while trying to simultaneously buy (or sell) contracts for a different expiration, one can enter into trades to execute the two sides in one trade, at a defined spread. Those price differences can be converted into percentage differences (and are shown in the columns to the right).

So, just as with outright futures, one can observe and/or enter a trade not to express a view that home price index levels will be higher in 2018 vs. 2017, but by how much. One can also trade one calendar spread versus another, but again the DEN contract already has priced in the highest gains (i.e. 3.8% on the bid side) so one would have to consider (in this instance) how much DEN will outperform another region.

I’ve posted 1×1 markets (one lot bid vs. one lot offered) to prompt discussion. I’d be open to facilitating larger-sized trades, and/or to providing quotes at levels inside those shown to further stimulate this conversation.

Please feel free to contact me (johnhdolan@homepricefutures.com) if you have any questions or trading axes.

Quarterly Pulsenomics home price survey

The results of 3rd quarterly home price survey by Pulsenomics have recently been released https://pulsenomics.com/Q3_2016_HPE_Survey.php. (Disclosure – I am a survey contributor).

I have a few observations on the numbers (as shown in table below):

Mean/Median expectations for 2016 are much higher. This may be a function of higher stock market, lower mortgage rates and low inventory levels. It is consistent with the stories from Denver, Portland, Seattle and other hot markets. While CME implied gains for 2016 have also been rising, 1) they reflect changes in Case Shiller index, not the Zillow index measured here, and 2) CME implied HPA (not shown here) is still lower than these survey results. (In comparing surveys vs. markets – particularly a market where interest seems dominated by hedgers – I wonder whether surveys, by their nature, are more optimistic, while markets where one can hedge might be quoted below “expected” future levels (if there’s an imbalance of interest).
Pulsenomic HPA gains are front-loaded, and with the passage of 3 months, have lower standard deviations, while longer term HPA projections are essentially unchanged, but with slightly wider projections. This seems consistent with recent comments in the press from economists (and the tone of political advertising?) expressing concern for the sustainability of home price growth.
Net, CME prices should translate into steeper calendar spreads on shorter expirations (which they have) and higher prices on longer-dated options (to reflect increase in standard deviation of expectations.)

Feel free to contact me (johnhdolan@homepricefutures.com) on this blog or any other aspect of hedging home prices.

Zillow forecast for Case Shiller CUS-10 ( in March)

While trading in longer dated contracts is more at the heart of hedging home price risks, my sense is that confidence in longer-dated markets is a function of believing that shorter-dated contracts are priced tight and correctly (that is to expectations of pending Case Shiller releases). Tight spot contract prices may also help feed, via debates over calendar spreads, into tighter longer-dated contract prices.

Thus I am a fan of anyone offering early, well-researched views on upcoming Case Shiller releases. (In case you were not aware) Zillow’s economists have been offering their predictions of the next month’s Case Shiller CUS-10 index for a long time. For example, this link shows how on Feb. 27- the day the Feb Case Shiller numbers were released -they offered their expectations for the March numbers.

Since March numbers are tallied from generally public data ending in January, it’s completely plausible that an entity with access to data, experts who can manage data, and researchers who focus on home price index calculations, should be able to come up with forecasts 3-4 weeks later. (Remember, the Case Shiller indices were conceived during a period pre-internet, when data collection and analysis took way more time. The two-month delay in releasing CS #’s made much more sense years ago.)

The Zillow MOM not-seasonally adjusted (NSA) prediction for March numbers is -0.1%. (Note that the headline Case Shiller numbers, as well as the CME contracts reference the not-seasonally adjusted series. Zillow also offers five other forecasts but the CUS-10 NSA forecast should be of particular interest to those trading CME futures.)

Not surprisingly CME prices for the May 2015 contract are not out of line versus Zillow’s prediction. Recent May contract quotes of 186.6/189.0 generate a mid-market value that is almost unchanged versus the February-released value of 187.81. Between some tail wind supporting home prices, and negative winter seasonal factors (which one can observe in Zillow’s seasonally adjusted forecasts) stable NSA Case Shiller values are reasonable.

So, please be aware of this Zillow offering. While one-month forecasts versus a contract trading with a wide bid/ask spread 3 months forward may have limited value today, pay close attention to Zillow’s April 28th forecast of May Case Shiller #’s. If history is a guide, by that point the May CME contract should be trading much tighter, and Zillow’s forecast will be on the settlement value of that contract.

As always, feel free to contact me (johnhdolan@homepricefutures.com) if you have any questions.

Dispersion of home price expectations

I continue to rely on (and participate in) the Pulsenomics quarterly survey on home price expectations (click here) both as a check against CME Case Shiller futures, and to identify economists with outlier calls. (“Come on guys. Back up your opinions with a trade”.) While the Pulsenomics survey uses the Zillow Home Price index, and the CME contracts reference Case Shiller indices, there’s a lot of parallels where knowledge about one product can be transferred to the other.

While I’d encourage you to go through the Pulsenomics link for the levels of home price expectations, I want to bring to your attention a trend in the survey results that should have good implications for trading in home price futures.

I’ve highlighted for months that bid/ask spreads in CME contracts have been compressing, and have noted that this should be a good precondition for trading, but some readers have observed that most of the contracts are quoted 1×1 (that is one lot bid, worth ~$50,000 on a contract priced at 200.0, versus one lot offered). How can readers get comfortable that the narrowing in bid/ask spreads might be indicative of wider forces?

The table to the right shows that while the level of home price expectations in the Pulsenomics survey over the last five quarters for 2017 and 2018 has been range-bound, the dispersion around those results has been narrowing rapidly. That supports the notion that there is greater consensus in expectations, less risk of a outlier event, and is therefore consistent with narrower bid/ask spreads in CME futures.

Pulsenomics conveys this theme in an even better graphic format here: (Focus on the black line)

Trading in CME contracts remains near historical lows, and I’m not aware of any OTC trades, but I would argue that the trend of declining standard deviation of home price expectations identified in the Pulsenomics survey should bode well for future trading opportunities.

As always, feel free to contact me (johnhdolan@homepricefutures.com) if you have any questions about this blog.

Aiming at Q14: Zillow forecasts vs bid/ask

Since so much of the CME Case Shiller futures trading takes place in the front contract, readers should be aware of any tools to help them in their projections of near-forward Case Shiller values. One of the best resources is the publicly available Zillow Real Estate Reserach blog http://www.zillow.com/research/april-2014-case-shiller-forecast-6997/ (link) where my sense is that their analysts have demonstrated a good track record in forecasting the next month CS index numbers.

Their call for the June release of the CUS 10-city index is a 1.9% month-on-month gain (NSA). That forecasts fits it neatly between the current spot index and bid/ask levels on the Q14 (Aug) CME contract (see graph below).

The graph shows the historical CS index values, bid/asks spreads on the Aug, Nov and Feb contracts, the mid-market levels, and my interpretation of the ZHI forecast for the June release.

Note: the Aug ’14 bid/ask are 3.6%/4.9% above spot levels.

As such, it would appear to me that some combination of the following factors must play out for CS index levels over the next 3 months:

Increases in the CS index levels will need to slow between June and Aug (despite ever more positive seasonals)
The Q14 contract levels will rise,
The ZHI is a tad too bullish, or
Recent slowdown in HPA will overwhelm strong seasonals resulting in a slowdown in June-Aug (release dates of April-June levels) CS index growth

I’m neutral on which of the four is most likely. I just invite people to read the Zillow work, and to the extent they disagree that they consider using the Q14 futures to express their views.

Zillow survey – Still many outliers

The quarterly Pulsenomics/Zillow home price expectations survey results were published last week. (See link). For those new to this space, this is probably the most comprehensive (and historic) survey of home price expectations. Over 100 contributors (including me) post their Q4/Q4 annual forecasts on a yearly basis out to 2018.

The survey results reference forecasts for the Zillow National Home Value Index (ZHVI) (click here for link to Zillow website) . While this is a change from past years when the Case Shiller National Home Price Index was the reference index, the results are similar- general bullishness with many outlier bulls and bears. The median estimated increase for the ZHVI index from 2012 Q4 to 2018 Q4 is ~27.3%, with a standard deviation of ~9.8%.

Both the Zillow and CS National index have fundamental differences versus the Case Shiller CUS 10-city index that the CME futures reference so any comparison between the survey results and CME futures needs to be qualified. As a aside, I posted a 25.2% increase for the Zillow Q4 2017 survey based on an observation of the CME CUS Nov ’17 mid-market quote that (adjusted for 3 additional months) was ~36.6.% above Dec. 2012 CS values. (For a discussion of the differences between ZHVI and the Case Shiller National index see the link in the Resources section or click here.)

No, my focus here is not converting one index to another, but with the outliers. Ten participants have 2018 index changes to be more greater than +40% (vs. 27.3% median), while seven are looking for gains of less than 15% (with two calling for negative 2012-2018 price changes). There are annual projected differences ranging from minus 5% in 2014 to 15% gains in 2016.

Even acknowledging the index differences, I call on these participants to either a) put their money where their mouth is, and/or b) to share with their followers that if they really believe these forecasts, that there is a platform (CME trading) where their readers can express these views, either in longer-term outright trades, or for those looking for outlier one-year results, in calendar spreads.

Feel free to contact me (johnhdolan@homepricefutures.com) if you want to learn more about how to convert your views into trades.

Pulsenomics/ Zillow home price survey

Today, Pulsenomics released the Zillow Home Price Expectations quarterly survey of >100 economists and real estate fans (including me), of home price forecasts for the next few years. Click here for link to survey results. This survey represents a break from prior ones in that the reference index for forecasts has switched from the Case Shiller National index to the US Zillow Home Value Index (ZHVI).

The change might result in small adjustments to forecasts. While the ZHVI and Case Shiller CUS10 index have “generally” moved in the same direction over the last few years, they diverged in the runup and fall in home prices during the 2005-2008 period. The CUS10 index also has seemed to show more variability over the last few years. ( I welcome someone to write a guest column highlighting the differences and possible implications for forward pricing.)

(Note that rather than contrast the Case Shiller National Index (which is released quarterly) with the ZHVI index (which is released monthly), the graph to the right shows the monthly ZHVI index versus the monthly CUS10 index. While this introduces yet another index, the CUS index is the reference for the CME-CUS futures contract so it may be more familiar to readers of this column.)

Net – the Zillow survey remains the most comprehensive survey of home price expectations that I am aware of. Anyone looking to trade housing derivatives should be aware of the results (headline +3% for 2013, but read the entire release for details and color), and as importantly, the dispersion in the forecasts.

I leave readers to click on the pulsenomics link to see the charts and tables. I’m going to do my quarterly badgering of the survey outliers to see if I can guilt them into a trade.

As always, please feel free to contact me johnhdolan@homepricefutures.com if you’d like to discuss the survey or any aspect of housing derivatives.

CME CS Futures vs. Expectations

Last week I reviewed how the CME Case Shiller futures lined up against the DB 3-year forecast on home prices. Today (Dec. 26) , the Pulsenomics/Zillow home price survey (for home prices out to 2017) was released. (click here.) This is the best survey of forward home price expectations, is printed quarterly, has a large sample size of over 100 real estate “experts” (full disclosure: including me), and should be on anyone’s radar who deals with home prices. Here’s a comparison of the results of that survey versus the CME markets.

First, there are some features of a comparison to the Zillow survey that existed in the comparison to the DB forecasts. Recall that the DB survey (as well as Zillow) was listing forecasts for changes at year-end, while the most liquid CME future is the November contract that references the 3-month period ending in September. Thus not only are the forecast periods slightly different, but the mismatchs introduce some seasonality that can’t be resolved with straight-line exptrapolation. Further, while the DB report was written a few days prior to the CME market prices I compared them to, the Zillow survey participants made their forecasts between late November and mid-December. As such expectations or markets might have changed.

Finally, the Zillow survey references the Case Shiller National Index, while the CME futures reference the CS 10-city index. The 10-city index covers the larger metropolitan regions (e.g. LAX, NYM) and as such may experience different price paths (both historical and going forward) than the National index.

Given all of those qualifiers the graph below highlights elements of the survey and the CME futures.

The black line is the historical CS CUS (10-city) index. The blue and red lines are the )Dec. 26) bid and offer lines for the CUS contracts that expire between Feb 2013 and Nov 2017. All three reference the left axis. The large blue squares represents the average of the Zillow survey, and the dashed lines above and below the squares represent +/- 1 standard deviation. Those numbers are scaled to the right axis. There is also a large brown square that represents the value for CUS-10 for December 2011.

The horizontal axis is the month in which the index is reported (so November for the later CME contracts and February for the year-end forecasts.

To illustrate, the CME Feb ’14 market (the one that settles on the Dec 2013 index) was 15880/16640 or 6.2%/11.2% above the Dec. 2011 value (149.59). (Recall that the CME trades at CS values *100). While the average forecast on the Zillow survey for home price appreciation from Q4 Dec. 2011 (7.87%) was inside this 6.2-11.2% range, more than 30% of the survey respondants were either more bullish than 11.8% (above the offered side of the CME contracts) or more bearish than 3.94% (well below the bid side of the CME contracts). It sounds like there should be some trades from these outlier bulls & bears (or at least from their clients.)

While the Feb 2013 and 2014 CME contracts reference the same year-end periods as the Zillow surveys, there are no Feburary contracts beyond 2014. The graph illustrates the timing differences.

Even given all of the qualifiers, one can observe that the Zilow averages tend to fall below the CME mid-market levels, often just about on top of the CME bids. Either the futures markets tend to be more bullish, OR, the differences between characteristics of the two indices are enough to justify the price differences.

If one believes that the larger CME markets are going to out-perform the National index, then relative shapes of the two graphs might make sense. But, if one believes that the regional markets (think Phoenix, Charolette, Seattle) are going to be stronger than NYM and LAX, then either Zillow expectations are too low, or futures prices are too high.

As with many discussions about home price forecasts, then, this discussion comes down to an understanding of the differences between the various indices being discussed. One can’t say that either the CME futures or the Zillow results are out of line, without having a view on now the Case Shiller National index is expected to perform relative to the 10-city index.

If you have ideas on that debate, please feel free to share them by contacting me at johnhdolan@homepricefutures.com.