SDG – HomePriceFutures.com

Moving? A potential application of Inter-city (IC) spreads

Redfin recently published a great piece that detailed how some residents look beyond their local area for their next house. They noted that many many people in a region look at the same next cities (sorted both by in-state and out-of-state searches) with certain marked preferences. For example, many LA residents eye San Diego, New Yorkers frequently consider Boston, and Denver homeowners seem to covet Seattle’s low taxes.

The graph below (from Redfin) show the net flow of users searching for other regions. (Net flow is the number looking to leave versus the number looking to take up residence). Seattle has recently re-taken the award for most people looking to move there, while Denver residents are looking elsewhere in larger numbers. This is interesting data to those who believe that population flows are a key component to changes in home prices.

Residents moving from one region another may face the risk that home prices will fall where they live, while rising where they want to move. Fortunately, my analysis suggests, and other research confirms, that home prices on many of the common pairs that Redfin identified are highly correlated, using region-wide measurements. ^1 For example, using YOY changes in Case Shiller indicies, the LAX, SDG and SFR regions have all been >90% correlated since 2013.^2,3 Surprisingly (to this former Connecticut resident now living in DC) the correlation across the three Northeast regions is not as strong, ranging from as low as 43% between NYM and WDC, to >75% between BOS and NYM and WDC.

For those in the later categories, intercity spread trades may be a useful tool for going simultaneously short futures on the region they’re leaving, while going long on the region to which they hope to move.^4 In effect, users might be able to hedge some of the risks of their move.

The table below lists (on the left side) the top 6 interstate 2018 transitions highlighted in the Redfin highlighted report. (I’ve added the intrastate move from LA to San Diego, as there are CME contracts on both). While the Redfin article has much more detail on the net number of people moving, my contribution to this discussion is to share how the CME markets (using the Feb ’20 contract to capture 2019 year-end values) might be used to possibly hedge these moves.

First, here’s a few explanations and observations:

For both the “From” and “To” cities, I’ve listed the current spot index, the bid, ask, and mid-market for the Feb ’20 (G20) contract on each city, or levels that I’d consider OTC trades (i.e. for Seattle, and Phoenix -highlighted in yellow)^5. The “Mid/ Spot -1” columns show the difference (in percentage terms) between the spot index for each city (both “From” and “To”) versus the spot index for each city.
Note that between the pairs of “From” and “To” cities, there are a total of 10 cities, and that for all but 3 (Seattle, Denver, and Phoenix) the “Mid/Spot -1” is a negative percent. Recall that the CME figures (at least for one-year forward contracts) don’t necessarily represent expectations of lower prices. The bids and offers are just levels that traders are willing to buy or sell contracts.
Note that in 6 of the 7 moves that Redfin highlighted, the “Mid/ Spot -1” value is higher in the destination city. (The difference between the two is shown in the “Diff” column.) Thus, it seems that either people may be moving to the more booming cities, or the act of people moving to those areas is correlated with better performing home contracts (as measured by the “Mid/Spot -1” metric.
Note that, the San Fran to Seattle move looks to have the biggest “payup” (i.e. going from a region where these quotes would be consistent with the SEX (Seattle) index outperforming the SFR index. By contrast, the move from Washington DC to New York, seems like the best bargain, as these quotes are consistent with WDC (Washington DC) index prices falling slower than NYM.

Finally, I’ve added levels where I’d be open to an intercity spread trade. For example in the NY to Boston move, the difference between the NYM and BOS “Mid/ Spot -1” numbers is 1.85%. Since both regions have a CME contract, I’d buy a NY contract while selling a BOS contract (a pair someone moving might be interested in ) where Boston outperforms NY by 3%, or go the reverse (buying Boston/ selling New York) where Boston outperforms by 1%. (See example below).
I’ve not posted this particular IC trade (or any of the others) as best orchestrated off-exchange, or users will need access to a broker who can execute IC trades.
While I’ve focused on the pairs Redfin highlighted, and for the Feb ’20 contract, in concept it’s possible to construct any other pairs of indices, for any other expiration.

Please feel free to contact me (johnhdolan@homepricefutures.com) if you’d like to discuss any aspect of this blog, want to move forward with any of the IC proposals discussed, or have questions on any aspect of hedging home price indices.

Thanks,

John

Footnotes:

^1 This might be no surprise as the people that like the attributes of one region seem to see similar attributes in another.

^2 Recall that correlation means that they move in the same direction at the same time, but not necessarily by the same amount/percent.

^3 Note that these are changes on the index, not any particular expiration of a futures contract.

^4 Even for those moving between two highly correlated regions, futures might be a useful tool should one be increasing/decreasing exposure.

^5 Contact me for details on how one might trade any home price index, not listed on the CME.

Majority of regions tipping toward negative HPA in 2019, falling from there in 2020.

Quotes on CME Case Shiller home price index futures have tipped over toward being consistent with negative HPA in 2019 for five of the ten regions (NYM, CHI, LAX, SDG, and SFR). Of the other fiver regions none are much above zero HPA, with BOS, DEN, and WDC just below 1%. Only BOS and LAV are priced for more than 1% gains against today’s spot levels.

The graph below is a candle graph that I use in monthly recaps. Price quotes on each of 11 CME regions (the 10-city index contract (HCI) and each of the ten components) have been converted into percent vs. spot for bid, ask and mid. For example, the BOSX19 bid, ask, and mid are 218.0, 221.0, and 219.5, while the BOS spot index is 216.56. Each bar thus shows the relative width of the bid/ask spread as well as the pricing vs. spot for each region, standardized on percentage terms. The outliers include LAV (priced consistent with gains for 2019), and SFR (priced at 2-3% declines through 2019 and 2020.) There are 22 bars, one for X19 (Nov ’19) and one for X20 (Nov ’20) for each of the 11 regions.

The graphs shows which contracts are above 100% (i.e. where prices would be unchanged versus spot), and compares the X19 to X20 ratios. Note that as a general rule, regions with lower prices for X19, tend to have even lower prices for X20. Hence, the net inference one can draw is that quotes are consistent with falling home price indices in 2019 for half the contracts, but that prices are lower in 2020 (vs. 2019) for almost all.

I’ve written recently how all CME quotes need to be taken with a huge grain of salt. The biggest point is the thinness of trading, as well as my belief (that I will explore in an upcoming blog) that longer-dated (i.e. more than 9 months) futures prices tend to be quoted at a discount to expectations. (See Dec 11 blog: Pulsenomics survey: Why are surveys results more bullish than Case Shiller futures? for more perspective). That said, several contracts are offered at discount to spot, or lower in 2020 than 2019, and no one’s “disagreed” (via an improved bid or purchase). The contracts need buyers to balance hedging inquiries I receive. Any takers willing to add what it likely a non-correlated risk to their corporate risk portfolio?

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

CME Markets post this morning’s Case Shiller #’s

Quotes on the CME Case Shiller home price index futures are generally lower by about 1/2 point per contract, after this morning’s release of the monthly Case Shiller numbers. The table below shows prices for the 11 Nov. 2019 contracts (X19) comparing yesterday versus today. The LAV and WDC contracts are higher (based on mid-markets) but all nine other contracts are lower. Home price indices on the coasts (BOS/NYM and LAX/SDG/SFR) faired worse than those inland. (This despite substantial upward revisions in last month’s NYM index.)

Forward curves continue to come under pressure. While the S&P press release led with the headline “Annual gains fall below 6% for the first time in 12 months”, year-on-year price differences on several contracts are consistent with HPA dropping to <2% by 2020.

There have been no trades yet today, but 22 contracts have traded this month, the largest monthly total since Jan. 2017. (My sense is that volume picks up when markets change direction. The biggest volumes were in the first months of the Financial Crises and in Spring 2012 when home price bottomed.)

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

Thanks, John

Revisiting ways to play the impact of a lower “Mortgage Interest Deduction” (MID

I had previously written about possible ways for market participants to play the impact of the prospects to changes in the Mortgage Interest Deduction (MID) in a blog on Oct 21 . Since this remains a timely topic, I again want to offer two opportunities for fans of the NAR’s (National Assoc. of Realtors) views, on how they might play the pending collapse in home prices in the higher- home price areas that have historically taken advantage of MID.

The first would be to enter either an OTC spread trade, or a total rate-of-return swap (TROR) on the difference between the performance of the Case Shiller 10-city index and the Case Shiller National index over some time frame. The CS-10 index contains many of the areas with higher-priced homes (e.g. NY, San Fran, Los Angeles) while the National index covers a much wider cross-section of the country that includes many lower-priced homes, where borrowers would presumably be less impacted by a cut in the ceiling in the amount of mortgage debt subject to the MID. The graphs to the right show the index values for the Case Shiller 10-city, 20-city and National indices (on top) as well as the year-over-year percentage changes in the indices.

Note that while the 10- and 20-city indices did much better in 2013, that the National index has recently been outperforming the 10-city index (i.e. 6.07% vs 5.33% for the last year, as of the October Case Shiller numbers released in Oct.) I would be open to an OTC trade on the differences between the index levels one-year forward (currently 216.5 on the 10-city index and 195.1 on the National index), or on the difference in percent gains (currently 74 bps) on the YOY gains in the National versus 10-city index.

While the above speaks to possible OTC trades, one can also take a more targeted view (by area) on existing CME products (i.e. Intercity Spreads). That is, one can “bet” on the performance of a particular regional index versus the Case Shiller 10-city index, if one truly believes that the areas with the highest priced homes will suffer relative price declines with the lowering of the cap on mortgage interest deduction.

The table above lists six regions that have both larger than average home values (using the Zillow ZHVI) as well as having regional CME Case Shiller futures contracts. (Note that while Zillow and Case Shiller geographic regions are not a perfect overlap, the ZHVI index does a good job of comparing the average prices to their national average. By any measure these six regions have higher-priced homes.)

In addition to a trader just outright selling the NYMX18 or SFRX18 contracts (X18= Nov 2018), they could also enter into an Intercity Spread trade where they simultaneously buy one contract and sell the other (in this case the CS 10-city index contract -HCI). This might be a more conservative play than just an outright sale as the end result will be a function of the difference between the two indices referenced in a trade. These IC contracts are listed where the bid side shows the price difference between where one might buy the HCI contract while selling the regional contract. So, for example, the -29.6 bid on HCI/SFR-X18, displays the bidder’s willingness to buy the HCIX18 contract at 224.8 while selling the SFRX18 contract at 254.4.

Note that the 224.4 price on the HCIX18 contract is 3.84% above the current spot level (of 216.49) while the SFRX18 price of 254.4 is 4.47% above the SFR spot index of 243.52. In effect, were one able to execute this IC trade on the bid side, one would be selling the SFRX index (for Nov 2018 settlement) with higher priced gains over the next 13 months, than the HCI index by 63 bps.

(The analysis in blue, shows the relative differences were a buyer to pay the offered side, in this case -28.0 points on the IC spread).

Note that if one were able to buy the IC spread on the bid side, they would be entering the trade at levels where in 5 of the 6 cases, the regional contract would be sold at a level with implied gains, higher than that of the regional contract. (The NYM contract is priced for lower price gains than the 10-city index).

Even if one bought the IC spread on the offered side, they would still be selling the BOS and DEN regional contract with higher priced gains than the HCI contract.

I am open to facilitating inquiries (or trades) on small amounts of such IC spreads to prompt further reaction to the MID debate. Please feel free to contact me (johnhdolan @homepricefutures.com) if you have any questions on this blog or any aspect of hedging home price indices.

Contrasting Pulsenomics Expectations vs. CME Forward Markets

I’d like to take the data and work done by two of my favorite sources (Pulsenomics and Getting Real) to further illustrate two points that I’ve been making here over the past few months.

First, as part of their quarterly survey of home price expectations across 100 participants (full disclosure, I’m one), Pulsenomics asked for opinions on how the home prices of 25 regions might perform in 2017. Their results are shown to the right. Please read their report for details on how they quantified relative price moves. For purposes of this illustration, higher (and positive is better and optimistic, while lower and/or negative is not).

Be aware that details of the regions identified (e.g. “Boston”), might be different than your understanding of that region, and important for later, the BOS Case Shiller index.

I then compared this tally versus the year-on- year gains for the these “regions” versus Case Shiller indices of the some name.

Observe (in the graph below) that the regions that have been identified in the Pulsenomics survey that are likely to do well in 2017, are typically the regions that did well in the last year (and vice versa). (FYI – 18 of the 25 regions used in the Pulsenomics survey have public Case Shiller indices. There are others, e.g. San Jose, but you have to subscribe to get them).

Real estate prices tend to have momentum, and so therefore this observation doesn’t surprise me, the current market for CME futures does not necessarily reflect the same high correlation.

First, note in the scatter diagram below that the forward prices for all 10 CME regional contracts (noted on the Y axis) falls below the red 45 degree line (with the slight exception of NYM). (FYI – To calculate the forward CME “prices” (in this graph) I took the mid-market values for the Nov 2018 contract and divided by the spot index (released in Nov). That percent gain is then converted back into annual gains). That is, forward HPA (as implied by CME prices) is much lower than gains over the last 12 months (measured with the X axis).

Second, there seems to be a reversion to the mean (in forward HPA) as forward gains for all contracts are converging to a narrow range (between +1.75-3.50%).

Third, away from any such reversion the slight outliers appear to be BOS (GE move?) and LAV (NFL/ beneficiary of construction?) while the below trend regions include CHI (pensions problems?) and the three California indices.

As I’ve noted (and conceded) in prior blogs, the California contracts seem out of line. While it could be that there are fundamental issues at work in California that I’ve not focused on. My contribution to this discussion is, that there may also be a seller who is larger than my potential interest in going long. In thinly traded markets, a small change in the balance between buyer and seller weight can have a disproportionate impact.

As I look to explain the California under-performance in the last few months, I can find lots of smaller hedgers that also want to sell. If there’s nothing “wrong” with the California real estate market, what do we need to do to entice prospective longs into dipping their toe in the longer-dated California contracts?

I’m happy to post any comments, share thoughts, or facilitate any trading ideas. Feel free to contact me (johnhdolan@homepricefutures.com) to discuss this blog or any other aspect of hedging home prices.

California falling, or an opportunity?

Since home prices turned up in 2012 the California markets (i.e. LAX, SDG and SFR indices) have consistently outperformed the CUS-10 index.

The graph (to the right) shows year-on-year percent changes in the CUS-index (in black) along with those for LAX, SDG and SFR. It has been a near truism for the last 4+ years that the rebound in California home prices has outperformed the CUS-10 index.

However, recently, pressure on longer-dated California CME futures prices has resulted in levels that are consistent with the California markets under-performing the CUS-10 index.

The chart below shows contract values for CUS-10 and the 3 California regions for the Nov ’17, ’18, ’19 and ’20 expirations translated into percent changes versus spot levels. (The height of each bar is the bid/ask levels and the mid-market value is shown with a green bar).

Note that the CUS bars are all higher than the California markets (barely so for LAX, but with obvious premiums over SDG and SFR). Net, the CUS-10 index is priced at levels that would reflect California index values under-performing over the next four years.

Now, as I noted in my monthly recap, this could be due to a variety of fundamental factors but I have two thoughts: 1) Given the lack of trading the contract prices have been weighed down by someone looking to sell, and 2) There is an opportunity to either buy the California markets at relatively (to the CUS 10 contract) cheap levels. In fact, if one doesn’t want to take outright risk the IC spread market (where one enters an order to simultaneously buy one contract (e.g. SDGX20) while selling another (e.g. HCIX20) at a predetermined spread, might be an attractive alternative.

Markets vary from fundamental values for a variety of reasons and order imbalance is often one (particularly in such a thinly traded market.)

Please feel free to contact me (johnhdolan@homepricefutures.com) if you’d like to discuss this theme, or any aspect of home price derivatives.

CME activity post June release of CS #’s

CME markets have been mostly quiet this morning as Case Shiller numbers came out generally in line toward CME quotes for the Aug ’15 contract. The two exceptions seem to have been in the BOS and SDG contracts where Nov ’15 contract prices have dropped sharply (as shown in the table below). There was one trade in SDGQ15 -the only one at the time of this report.

On the other hand, LAX and DEN prices moved higher as upward price momentum (at least in the non-seasonally adjusted numbers) continued.

Bid/ask spreads have recovered to yesterday’s levels after wider quotes before and after the 9:15 release. While spreads are about the same across all regions and all expirations today, on the month bid/ask spreads are much tighter -particularly in the CUS and CHI contracts, as well as the Q15 and X17 expirations. (I am traveling the next few days so I may not get the June recap out until during the July 4th weekend. Please contact me if you’d like to see a specific table earlier.)

As today is month-end expect some filling in of prices on less-frequently quoted contracts and a sharp pencil being taken to some calendar spreads. Any help would be appreciated as the tighter the month-end quotes, the better the contracts look when I go to market them.

If you have any questions please contact me @ johnhdolan@homepricefutures.com . I’m hoping that my tech and email problems are behind me as I’ve transitioned to a new server.

As Nov. expiration approaches

The last day of trading for the Nov 2013 contract is Monday. (Recall that the Case Shiller indices will be released on Tuesday and that the Nov 2013 contract will settle on those numbers.) With only a few days to go I thought that I’d tweak quotes and post a market status.

The attached table shows bid. ask, and mid-market levels for the 11 CME futures contracts (from earlier this morning). All bid-ask spreads are <= 2.0 points, although none is tighter than 1.4 points. In past expirations we’ve often seen one contract where value is being more hotly “debated” with a tighter bid/ask spread. Given that open interest for Nov. is 81 contracts (including 46 in CHI) there may be some traders looking to unwind (or roll forward) Nov positions. I’d expect some outright and short calendar spread trading over the next few days. The Nov /Feb calendar spread markets are wide at this point (not shown) probably due to the combination of changing seasonal factors (warm vs. cold states) and concerns about the underlying momentum in home prices. (Look for a blog on Nov /Feb spreads in the next few days.)

The Nov 2013 mid-market contract values are consistent with news headlines touting continued (but slowing) gains in all regional indices (at least those traded on the CME). The LAV, SDG and SFR will lead the year-over-year increases, while the Northeast areas of BOS, NYM and WDC will show as the laggards.

Finally, note that with the expiration of the Nov 2013 contract, the CME will introduce a new contract for Nov 2018. I would expect that most contracts roll out with 2-5% implied HPA gains over mid-2017 levels. With gains implied by such HPA, the forward level of CS indices will move even closer to full recovery versus 2006 “highs” in selected key regions. I would expect some push-back, increased interest in hedging as we get to “back to past high levels”.

Basics: What If CUS Index composition weights change?

The weighting for Case Shiller composite indices (e.g. CUS 10-city, referenced in the CME contracts) may be adjusted 2- 3 years after the 2010 census, per the Case Shiller index methodology. I raise this notion to: 1) increase awareness, 2) minimize concern, and 3) invite those who better understand index calculations to weigh in. I have no idea if, or when, this might happen, but as it might have an impact on CUS values, I wanted to start a (more public) discussion.

As a test, I used the changes in population for loosely defined geographic regions to try and estimate what the new weights might look like and how those changes might impact CUS values. I made the simplistic leap of faith that population growth is correlated to housing stock, which I believe, will be the denominator for assigning weights by region. The analysis is not intended to be precise but to illustrate some key moving parts in how a change in weighting might impact (if at all) the CUS index.

The population gains are primarily in the West (DEN, LAV, LAX, SDG and SFR) and South (MIA) while the losses are in the North (BOS, CHI, NYM). The WDC region bucks the Northeast trend. Given my “What If” weights, the CUS index would be impacted in this analysis by -0.18.

The reason that an adjustment (using my “What If” weights) is so small is that weighting is taken away from NYM (which trades at a fairly close level to the CUS index) and redistributed to both the low priced index areas of LAV and DEN (which would pull down the CUS value) AND the high-priced areas of SDG and WDC.

(Note that any change in the allocation to SFR should have a small impact on spot CUS values as both indices have about the same value. However this (a change in redistribution to SFR) would not be neutral to longer-dated contracts where SFR contracts trades at a premium to CUS.)

Key to any estimation/conclusion is how future index weightings might vary from this simplistic analysis. Changes from/to the low index areas (DEN, LAV) to the high-priced index areas (LAX, WDC) might result in more pronounced changes to CUS.

Feel free to contact me (johnhdolan@homepricefutures.com) if you care to discuss this theme, or any other aspect of trading home price indices.